cuDF API Reference¶
DataFrame¶
-
class
cudf.core.dataframe.DataFrame(data=None, index=None, columns=None, dtype=None)¶ A GPU Dataframe object.
- Parameters
- dataarray-like, Iterable, dict, or DataFrame.
Dict can contain Series, arrays, constants, or list-like objects.
- indexIndex or array-like
Index to use for resulting frame. Will default to RangeIndex if no indexing information part of input data and no index provided.
- columnsIndex or array-like
Column labels to use for resulting frame. Will default to RangeIndex (0, 1, 2, …, n) if no column labels are provided.
- dtypedtype, default None
Data type to force. Only a single dtype is allowed. If None, infer.
Examples
Build dataframe with
__setitem__:>>> import cudf >>> df = cudf.DataFrame() >>> df['key'] = [0, 1, 2, 3, 4] >>> df['val'] = [float(i + 10) for i in range(5)] # insert column >>> df key val 0 0 10.0 1 1 11.0 2 2 12.0 3 3 13.0 4 4 14.0
Build DataFrame via dict of columns:
>>> import numpy as np >>> from datetime import datetime, timedelta >>> t0 = datetime.strptime('2018-10-07 12:00:00', '%Y-%m-%d %H:%M:%S') >>> n = 5 >>> df = cudf.DataFrame({ ... 'id': np.arange(n), ... 'datetimes': np.array( ... [(t0+ timedelta(seconds=x)) for x in range(n)]) ... }) >>> df id datetimes 0 0 2018-10-07T12:00:00.000 1 1 2018-10-07T12:00:01.000 2 2 2018-10-07T12:00:02.000 3 3 2018-10-07T12:00:03.000 4 4 2018-10-07T12:00:04.000
Build DataFrame via list of rows as tuples:
>>> df = cudf.DataFrame([ ... (5, "cats", "jump", np.nan), ... (2, "dogs", "dig", 7.5), ... (3, "cows", "moo", -2.1, "occasionally"), ... ]) >>> df 0 1 2 3 4 0 5 cats jump <NA> <NA> 1 2 dogs dig 7.5 <NA> 2 3 cows moo -2.1 occasionally
Convert from a Pandas DataFrame:
>>> import pandas as pd >>> pdf = pd.DataFrame({'a': [0, 1, 2, 3],'b': [0.1, 0.2, None, 0.3]}) >>> pdf a b 0 0 0.1 1 1 0.2 2 2 NaN 3 3 0.3 >>> df = cudf.from_pandas(pdf) >>> df a b 0 0 0.1 1 1 0.2 2 2 <NA> 3 3 0.3
- Attributes
TTranspose index and columns.
atAlias for
DataFrame.loc; provided for compatibility with Pandas.columnsReturns a tuple of columns
dtypesReturn the dtypes in this object.
emptyIndicator whether DataFrame or Series is empty.
iatAlias for
DataFrame.iloc; provided for compatibility with Pandas.ilocSelecting rows and column by position.
indexReturns the index of the DataFrame
locSelecting rows and columns by label or boolean mask.
ndimDimension of the data.
shapeReturns a tuple representing the dimensionality of the DataFrame.
sizeReturn the number of elements in the underlying data.
valuesReturn a CuPy representation of the DataFrame.
Methods
acos()Get Trigonometric inverse cosine, element-wise.
add(other[, axis, level, fill_value])Get Addition of dataframe and other, element-wise (binary operator add).
agg(aggs[, axis])Aggregate using one or more operations over the specified axis.
all([axis, bool_only, skipna, level])Return whether all elements are True in DataFrame.
any([axis, bool_only, skipna, level])Return whether any elements is True in DataFrame.
append(other[, ignore_index, …])Append rows of other to the end of caller, returning a new object.
apply_chunks(func, incols, outcols[, …])Transform user-specified chunks using the user-provided function.
apply_rows(func, incols, outcols, kwargs[, …])Apply a row-wise user defined function.
argsort([ascending, na_position])Sort by the values.
as_gpu_matrix([columns, order])Convert to a matrix in device memory.
as_matrix([columns])Convert to a matrix in host memory.
asin()Get Trigonometric inverse sine, element-wise.
assign(**kwargs)Assign columns to DataFrame from keyword arguments.
astype(dtype[, copy, errors])Cast the DataFrame to the given dtype
atan()Get Trigonometric inverse tangent, element-wise.
clip([lower, upper, inplace, axis])Trim values at input threshold(s).
copy([deep])Make a copy of this object’s indices and data.
corr()Compute the correlation matrix of a DataFrame.
cos()Get Trigonometric cosine, element-wise.
count([axis, level, numeric_only])Count
non-NAcells for each column or row.cov(**kwargs)Compute the covariance matrix of a DataFrame.
cummax([axis, skipna])Return cumulative maximum of the DataFrame.
cummin([axis, skipna])Return cumulative minimum of the DataFrame.
cumprod([axis, skipna])Return cumulative product of the DataFrame.
cumsum([axis, skipna])Return cumulative sum of the DataFrame.
describe([percentiles, include, exclude, …])Generate descriptive statistics.
div(other[, axis, level, fill_value])Get Floating division of dataframe and other, element-wise (binary operator truediv).
drop([labels, axis, index, columns, level, …])Drop specified labels from rows or columns.
drop_duplicates([subset, keep, inplace, …])Return DataFrame with duplicate rows removed, optionally only considering certain subset of columns.
dropna([axis, how, thresh, subset, inplace])Drops rows (or columns) containing nulls from a Column.
equals(other)Test whether two objects contain the same elements.
exp()Get the exponential of all elements, element-wise.
fillna([value, method, axis, inplace, limit])Fill null values with
valueor specifiedmethod.floordiv(other[, axis, level, fill_value])Get Integer division of dataframe and other, element-wise (binary operator floordiv).
from_arrow(table)Convert from PyArrow Table to DataFrame.
from_pandas(dataframe[, nan_as_null])Convert from a Pandas DataFrame.
from_records(data[, index, columns, nan_as_null])Convert structured or record ndarray to DataFrame.
hash_columns([columns])Hash the given columns and return a new device array
head([n])Returns the first n rows as a new DataFrame
info([verbose, buf, max_cols, memory_usage, …])Print a concise summary of a DataFrame.
insert(loc, name, value)Add a column to DataFrame at the index specified by loc.
Interleave Series columns of a table into a single column.
isin(values)Whether each element in the DataFrame is contained in values.
isna()Identify missing values.
isnull()Identify missing values.
Iterate over column names and series pairs
join(other[, on, how, lsuffix, rsuffix, …])Join columns with other DataFrame on index or on a key column.
keys()Get the columns.
kurt([axis, skipna, level, numeric_only])Return Fisher’s unbiased kurtosis of a sample.
kurtosis([axis, skipna, level, numeric_only])Return Fisher’s unbiased kurtosis of a sample.
label_encoding(column, prefix, cats[, …])Encode labels in a column with label encoding.
log()Get the natural logarithm of all elements, element-wise.
mask(cond[, other, inplace])Replace values where the condition is True.
max([axis, skipna, level, numeric_only])Return the maximum of the values in the DataFrame.
mean([axis, skipna, level, numeric_only])Return the mean of the values for the requested axis.
melt(**kwargs)Unpivots a DataFrame from wide format to long format, optionally leaving identifier variables set.
memory_usage([index, deep])Return the memory usage of each column in bytes.
merge(right[, on, left_on, right_on, …])Merge GPU DataFrame objects by performing a database-style join operation by columns or indexes.
min([axis, skipna, level, numeric_only])Return the minimum of the values in the DataFrame.
mod(other[, axis, level, fill_value])Get Modulo division of dataframe and other, element-wise (binary operator mod).
mode([axis, numeric_only, dropna])Get the mode(s) of each element along the selected axis.
mul(other[, axis, level, fill_value])Get Multiplication of dataframe and other, element-wise (binary operator mul).
Convert nans (if any) to nulls.
nlargest(n, columns[, keep])Get the rows of the DataFrame sorted by the n largest value of columns
notna()Identify non-missing values.
notnull()Identify non-missing values.
nsmallest(n, columns[, keep])Get the rows of the DataFrame sorted by the n smallest value of columns
one_hot_encoding(column, prefix, cats[, …])Expand a column with one-hot-encoding.
partition_by_hash(columns, nparts[, keep_index])Partition the dataframe by the hashed value of data in columns.
pipe(func, *args, **kwargs)Apply
func(self, *args, **kwargs).pivot(index, columns[, values])Return reshaped DataFrame organized by the given index and column values.
pop(item)Return a column and drop it from the DataFrame.
pow(other[, axis, level, fill_value])Get Exponential power of dataframe and other, element-wise (binary operator pow).
prod([axis, skipna, dtype, level, …])Return product of the values in the DataFrame.
product([axis, skipna, dtype, level, …])Return product of the values in the DataFrame.
quantile([q, axis, numeric_only, …])Return values at the given quantile.
quantiles([q, interpolation])Return values at the given quantile.
query(expr[, local_dict])Query with a boolean expression using Numba to compile a GPU kernel.
radd(other[, axis, level, fill_value])Get Addition of dataframe and other, element-wise (binary operator radd).
rank([axis, method, numeric_only, …])Compute numerical data ranks (1 through n) along axis.
rdiv(other[, axis, level, fill_value])Get Floating division of dataframe and other, element-wise (binary operator rtruediv).
reindex([labels, axis, index, columns, copy])Return a new DataFrame whose axes conform to a new index
rename([mapper, index, columns, axis, copy, …])Alter column and index labels.
repeat(repeats[, axis])Repeats elements consecutively.
replace([to_replace, value, inplace, limit, …])Replace values given in to_replace with replacement.
reset_index([level, drop, inplace, …])Reset the index.
rfloordiv(other[, axis, level, fill_value])Get Integer division of dataframe and other, element-wise (binary operator rfloordiv).
rmod(other[, axis, level, fill_value])Get Modulo division of dataframe and other, element-wise (binary operator rmod).
rmul(other[, axis, level, fill_value])Get Multiplication of dataframe and other, element-wise (binary operator rmul).
round([decimals])Round a DataFrame to a variable number of decimal places.
rpow(other[, axis, level, fill_value])Get Exponential power of dataframe and other, element-wise (binary operator pow).
rsub(other[, axis, level, fill_value])Get Subtraction of dataframe and other, element-wise (binary operator rsub).
rtruediv(other[, axis, level, fill_value])Get Floating division of dataframe and other, element-wise (binary operator rtruediv).
sample([n, frac, replace, weights, …])Return a random sample of items from an axis of object.
scatter_by_map(map_index[, map_size, keep_index])Scatter to a list of dataframes.
searchsorted(values[, side, ascending, …])Find indices where elements should be inserted to maintain order
select_dtypes([include, exclude])Return a subset of the DataFrame’s columns based on the column dtypes.
set_index(keys[, drop, append, inplace, …])Return a new DataFrame with a new index
shift([periods, freq, axis, fill_value])Shift values by periods positions.
sin()Get Trigonometric sine, element-wise.
skew([axis, skipna, level, numeric_only])Return unbiased Fisher-Pearson skew of a sample.
sort_index([axis, level, ascending, …])Sort object by labels (along an axis).
sort_values(by[, axis, ascending, inplace, …])Sort by the values row-wise.
sqrt()Get the non-negative square-root of all elements, element-wise.
stack([level, dropna])Stack the prescribed level(s) from columns to index
std([axis, skipna, level, ddof, numeric_only])Return sample standard deviation of the DataFrame.
sub(other[, axis, level, fill_value])Get Subtraction of dataframe and other, element-wise (binary operator sub).
sum([axis, skipna, dtype, level, …])Return sum of the values in the DataFrame.
tail([n])Returns the last n rows as a new DataFrame
take(positions[, keep_index])Return a new DataFrame containing the rows specified by positions
tan()Get Trigonometric tangent, element-wise.
tile(count)Repeats the rows from self DataFrame count times to form a new DataFrame.
to_arrow([preserve_index])Convert to a PyArrow Table.
to_csv([path_or_buf, sep, na_rep, columns, …])Write a dataframe to csv file format.
Converts a cuDF object into a DLPack tensor.
to_feather(path, *args, **kwargs)Write a DataFrame to the feather format.
to_hdf(path_or_buf, key, *args, **kwargs)Write the contained data to an HDF5 file using HDFStore.
to_json([path_or_buf])Convert the cuDF object to a JSON string.
to_orc(fname[, compression])Write a DataFrame to the ORC format.
to_pandas([nullable])Convert to a Pandas DataFrame.
to_parquet(path, *args, **kwargs)Write a DataFrame to the parquet format.
to_records([index])Convert to a numpy recarray
Convert to string
Transpose index and columns.
truediv(other[, axis, level, fill_value])Get Floating division of dataframe and other, element-wise (binary operator truediv).
unstack([level, fill_value])Pivot one or more levels of the (necessarily hierarchical) index labels.
update(other[, join, overwrite, …])Modify a DataFrame in place using non-NA values from another DataFrame.
var([axis, skipna, level, ddof, numeric_only])Return unbiased variance of the DataFrame.
where(cond[, other, inplace])Replace values where the condition is False.
groupby
rolling
-
property
T¶ Transpose index and columns.
Reflect the DataFrame over its main diagonal by writing rows as columns and vice-versa. The property T is an accessor to the method transpose().
- Returns
- outDataFrame
The transposed DataFrame.
-
acos()¶ Get Trigonometric inverse cosine, element-wise.
The inverse of cos so that, if y = x.cos(), then x = y.acos()
- Returns
- DataFrame/Series/Index
Result of the trigonometric operation.
Examples
>>> import cudf >>> ser = cudf.Series([-1, 0, 1, 0.32434, 0.5]) >>> ser.acos() 0 3.141593 1 1.570796 2 0.000000 3 1.240482 4 1.047198 dtype: float64
acos operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-1, 0, 0.5], ... 'second': [0.234, 0.3, 0.1]}) >>> df first second 0 -1.0 0.234 1 0.0 0.300 2 0.5 0.100 >>> df.acos() first second 0 3.141593 1.334606 1 1.570796 1.266104 2 1.047198 1.470629
acos operation on Index:
>>> index = cudf.Index([-1, 0.4, 1, 0, 0.3]) >>> index Float64Index([-1.0, 0.4, 1.0, 0.0, 0.3], dtype='float64') >>> index.acos() Float64Index([ 3.141592653589793, 1.1592794807274085, 0.0, 1.5707963267948966, 1.266103672779499], dtype='float64')
-
add(other, axis='columns', level=None, fill_value=None)¶ Get Addition of dataframe and other, element-wise (binary operator add).
Equivalent to
dataframe + other, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, radd.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df + 1 angles degrees circle 1 361 triangle 4 181 rectangle 5 361 >>> df.add(1) angles degrees circle 1 361 triangle 4 181 rectangle 5 361
-
agg(aggs, axis=None)¶ Aggregate using one or more operations over the specified axis.
- Parameters
- aggsIterable (set, list, string, tuple or dict)
- Function to use for aggregating data. Accepted types are:
string name, e.g.
"sum"list of functions, e.g.
["sum", "min", "max"]dict of axis labels specified operations per column, e.g.
{"a": "sum"}
- axisnot yet supported
- Returns
- Aggregation Result
SeriesorDataFrame When
DataFrame.aggis called with single agg,Seriesis returned. WhenDataFrame.aggis called with several aggs,DataFrameis returned.
- Aggregation Result
Notes
- Difference from pandas:
Not supporting:
axis,*args,**kwargs
-
all(axis=0, bool_only=None, skipna=True, level=None, **kwargs)¶ Return whether all elements are True in DataFrame.
- Parameters
- skipna: bool, default True
Exclude NA/null values. If the entire row/column is NA and skipna is True, then the result will be True, as for an empty row/column. If skipna is False, then NA are treated as True, because these are not equal to zero.
- Returns
- Series
Notes
Parameters currently not supported are axis, bool_only, level.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [3, 2, 3, 4], 'b': [7, 0, 10, 10]}) >>> df.all() a True b False dtype: bool
-
any(axis=0, bool_only=None, skipna=True, level=None, **kwargs)¶ Return whether any elements is True in DataFrame.
- Parameters
- skipna: bool, default True
Exclude NA/null values. If the entire row/column is NA and skipna is True, then the result will be False, as for an empty row/column. If skipna is False, then NA are treated as True, because these are not equal to zero.
- Returns
- Series
Notes
Parameters currently not supported are axis, bool_only, level.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [3, 2, 3, 4], 'b': [7, 0, 10, 10]}) >>> df.any() a True b True dtype: bool
-
append(other, ignore_index=False, verify_integrity=False, sort=False)¶ Append rows of other to the end of caller, returning a new object. Columns in other that are not in the caller are added as new columns.
- Parameters
- otherDataFrame or Series/dict-like object, or list of these
The data to append.
- ignore_indexbool, default False
If True, do not use the index labels.
- sortbool, default False
Sort columns ordering if the columns of self and other are not aligned.
- verify_integritybool, default False
This Parameter is currently not supported.
- Returns
- DataFrame
See also
cudf.core.reshape.concatGeneral function to concatenate DataFrame or objects.
Notes
If a list of dict/series is passed and the keys are all contained in the DataFrame’s index, the order of the columns in the resulting DataFrame will be unchanged. Iteratively appending rows to a cudf DataFrame can be more computationally intensive than a single concatenate. A better solution is to append those rows to a list and then concatenate the list with the original DataFrame all at once. verify_integrity parameter is not supported yet.
Examples
>>> import cudf >>> df = cudf.DataFrame([[1, 2], [3, 4]], columns=list('AB')) >>> df A B 0 1 2 1 3 4 >>> df2 = cudf.DataFrame([[5, 6], [7, 8]], columns=list('AB')) >>> df2 A B 0 5 6 1 7 8 >>> df.append(df2) A B 0 1 2 1 3 4 0 5 6 1 7 8
With ignore_index set to True:
>>> df.append(df2, ignore_index=True) A B 0 1 2 1 3 4 2 5 6 3 7 8
The following, while not recommended methods for generating DataFrames, show two ways to generate a DataFrame from multiple data sources. Less efficient:
>>> df = cudf.DataFrame(columns=['A']) >>> for i in range(5): ... df = df.append({'A': i}, ignore_index=True) >>> df A 0 0 1 1 2 2 3 3 4 4
More efficient than above:
>>> cudf.concat([cudf.DataFrame([i], columns=['A']) for i in range(5)], ... ignore_index=True) A 0 0 1 1 2 2 3 3 4 4
-
apply_chunks(func, incols, outcols, kwargs=None, pessimistic_nulls=True, chunks=None, blkct=None, tpb=None)¶ Transform user-specified chunks using the user-provided function.
- Parameters
- dfDataFrame
The source dataframe.
- funcfunction
The transformation function that will be executed on the CUDA GPU.
- incols: list or dict
A list of names of input columns that match the function arguments. Or, a dictionary mapping input column names to their corresponding function arguments such as {‘col1’: ‘arg1’}.
- outcols: dict
A dictionary of output column names and their dtype.
- kwargs: dict
name-value of extra arguments. These values are passed directly into the function.
- pessimistic_nullsbool
Whether or not apply_rows output should be null when any corresponding input is null. If False, all outputs will be non-null, but will be the result of applying func against the underlying column data, which may be garbage.
- chunksint or Series-like
If it is an
int, it is the chunksize. If it is an array, it contains integer offset for the start of each chunk. The span of a chunk for chunk i-th isdata[chunks[i] : chunks[i + 1]]for anyi + 1 < chunks.size; or,data[chunks[i]:]for thei == len(chunks) - 1.- tpbint; optional
The threads-per-block for the underlying kernel. If not specified (Default), uses Numba
.forall(...)built-in to query the CUDA Driver API to determine optimal kernel launch configuration. Specify 1 to emulate serial execution for each chunk. It is a good starting point but inefficient. Its maximum possible value is limited by the available CUDA GPU resources.- blkctint; optional
The number of blocks for the underlying kernel. If not specified (Default) and
tpbis not specified (Default), uses Numba.forall(...)built-in to query the CUDA Driver API to determine optimal kernel launch configuration. If not specified (Default) andtpbis specified, useschunksas the number of blocks.
See also
Examples
For
tpb > 1,funcis executed bytpbnumber of threads concurrently. To access the thread id and count, usenumba.cuda.threadIdx.xandnumba.cuda.blockDim.x, respectively (See numba CUDA kernel documentation).In the example below, the kernel is invoked concurrently on each specified chunk. The kernel computes the corresponding output for the chunk.
By looping over the range
range(cuda.threadIdx.x, in1.size, cuda.blockDim.x), the kernel function can be used with any tpb in an efficient manner.>>> from numba import cuda >>> @cuda.jit ... def kernel(in1, in2, in3, out1): ... for i in range(cuda.threadIdx.x, in1.size, cuda.blockDim.x): ... x = in1[i] ... y = in2[i] ... z = in3[i] ... out1[i] = x * y + z
-
apply_rows(func, incols, outcols, kwargs, pessimistic_nulls=True, cache_key=None)¶ Apply a row-wise user defined function.
- Parameters
- dfDataFrame
The source dataframe.
- funcfunction
The transformation function that will be executed on the CUDA GPU.
- incols: list or dict
A list of names of input columns that match the function arguments. Or, a dictionary mapping input column names to their corresponding function arguments such as {‘col1’: ‘arg1’}.
- outcols: dict
A dictionary of output column names and their dtype.
- kwargs: dict
name-value of extra arguments. These values are passed directly into the function.
- pessimistic_nullsbool
Whether or not apply_rows output should be null when any corresponding input is null. If False, all outputs will be non-null, but will be the result of applying func against the underlying column data, which may be garbage.
Examples
The user function should loop over the columns and set the output for each row. Loop execution order is arbitrary, so each iteration of the loop MUST be independent of each other.
When
funcis invoked, the array args corresponding to the input/output are strided so as to improve GPU parallelism. The loop in the function resembles serial code, but executes concurrently in multiple threads.>>> import cudf >>> import numpy as np >>> df = cudf.DataFrame() >>> nelem = 3 >>> df['in1'] = np.arange(nelem) >>> df['in2'] = np.arange(nelem) >>> df['in3'] = np.arange(nelem)
Define input columns for the kernel
>>> in1 = df['in1'] >>> in2 = df['in2'] >>> in3 = df['in3'] >>> def kernel(in1, in2, in3, out1, out2, kwarg1, kwarg2): ... for i, (x, y, z) in enumerate(zip(in1, in2, in3)): ... out1[i] = kwarg2 * x - kwarg1 * y ... out2[i] = y - kwarg1 * z
Call
.apply_rowswith the name of the input columns, the name and dtype of the output columns, and, optionally, a dict of extra arguments.>>> df.apply_rows(kernel, ... incols=['in1', 'in2', 'in3'], ... outcols=dict(out1=np.float64, out2=np.float64), ... kwargs=dict(kwarg1=3, kwarg2=4)) in1 in2 in3 out1 out2 0 0 0 0 0.0 0.0 1 1 1 1 1.0 -2.0 2 2 2 2 2.0 -4.0
-
argsort(ascending=True, na_position='last')¶ Sort by the values.
- Parameters
- ascendingbool or list of bool, default True
If True, sort values in ascending order, otherwise descending.
- na_position{‘first’ or ‘last’}, default ‘last’
Argument ‘first’ puts NaNs at the beginning, ‘last’ puts NaNs at the end.
- Returns
- out_column_indscuDF Column of indices sorted based on input
Notes
Difference from pandas:
Support axis=’index’ only.
Not supporting: inplace, kind
Ascending can be a list of bools to control per column
-
as_gpu_matrix(columns=None, order='F')¶ Convert to a matrix in device memory.
- Parameters
- columnssequence of str
List of a column names to be extracted. The order is preserved. If None is specified, all columns are used.
- order‘F’ or ‘C’
Optional argument to determine whether to return a column major (Fortran) matrix or a row major (C) matrix.
- Returns
- A (nrow x ncol) numba device ndarray
-
as_matrix(columns=None)¶ Convert to a matrix in host memory.
- Parameters
- columnssequence of str
List of a column names to be extracted. The order is preserved. If None is specified, all columns are used.
- Returns
- A (nrow x ncol) numpy ndarray in “F” order.
-
asin()¶ Get Trigonometric inverse sine, element-wise.
The inverse of sine so that, if y = x.sin(), then x = y.asin()
- Returns
- DataFrame/Series/Index
Result of the trigonometric operation.
Examples
>>> import cudf >>> ser = cudf.Series([-1, 0, 1, 0.32434, 0.5]) >>> ser.asin() 0 -1.570796 1 0.000000 2 1.570796 3 0.330314 4 0.523599 dtype: float64
asin operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-1, 0, 0.5], ... 'second': [0.234, 0.3, 0.1]}) >>> df first second 0 -1.0 0.234 1 0.0 0.300 2 0.5 0.100 >>> df.asin() first second 0 -1.570796 0.236190 1 0.000000 0.304693 2 0.523599 0.100167
asin operation on Index:
>>> index = cudf.Index([-1, 0.4, 1, 0.3]) >>> index Float64Index([-1.0, 0.4, 1.0, 0.3], dtype='float64') >>> index.asin() Float64Index([-1.5707963267948966, 0.41151684606748806, 1.5707963267948966, 0.3046926540153975], dtype='float64')
-
assign(**kwargs)¶ Assign columns to DataFrame from keyword arguments.
Examples
>>> import cudf >>> df = cudf.DataFrame() >>> df = df.assign(a=[0, 1, 2], b=[3, 4, 5]) >>> df a b 0 0 3 1 1 4 2 2 5
-
astype(dtype, copy=False, errors='raise', **kwargs)¶ Cast the DataFrame to the given dtype
- Parameters
- dtypedata type, or dict of column name -> data type
Use a numpy.dtype or Python type to cast entire DataFrame object to the same type. Alternatively, use
{col: dtype, ...}, where col is a column label and dtype is a numpy.dtype or Python type to cast one or more of the DataFrame’s columns to column-specific types.- copybool, default False
Return a deep-copy when
copy=True. Note by defaultcopy=Falsesetting is used and hence changes to values then may propagate to other cudf objects.- errors{‘raise’, ‘ignore’, ‘warn’}, default ‘raise’
Control raising of exceptions on invalid data for provided dtype.
raise: allow exceptions to be raisedignore: suppress exceptions. On error return original object.warn: prints last exceptions as warnings and return original object.
- **kwargsextra arguments to pass on to the constructor
- Returns
- castedDataFrame
-
property
at¶ Alias for
DataFrame.loc; provided for compatibility with Pandas.
-
atan()¶ Get Trigonometric inverse tangent, element-wise.
The inverse of tan so that, if y = x.tan(), then x = y.atan()
- Returns
- DataFrame/Series/Index
Result of the trigonometric operation.
Examples
>>> import cudf >>> ser = cudf.Series([-1, 0, 1, 0.32434, 0.5, -10]) >>> ser 0 -1.00000 1 0.00000 2 1.00000 3 0.32434 4 0.50000 5 -10.00000 dtype: float64 >>> ser.atan() 0 -0.785398 1 0.000000 2 0.785398 3 0.313635 4 0.463648 5 -1.471128 dtype: float64
atan operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-1, -10, 0.5], ... 'second': [0.234, 0.3, 10]}) >>> df first second 0 -1.0 0.234 1 -10.0 0.300 2 0.5 10.000 >>> df.atan() first second 0 -0.785398 0.229864 1 -1.471128 0.291457 2 0.463648 1.471128
atan operation on Index:
>>> index = cudf.Index([-1, 0.4, 1, 0, 0.3]) >>> index Float64Index([-1.0, 0.4, 1.0, 0.0, 0.3], dtype='float64') >>> index.atan() Float64Index([-0.7853981633974483, 0.3805063771123649, 0.7853981633974483, 0.0, 0.2914567944778671], dtype='float64')
-
clip(lower=None, upper=None, inplace=False, axis=1)¶ Trim values at input threshold(s).
Assigns values outside boundary to boundary values. Thresholds can be singular values or array like, and in the latter case the clipping is performed element-wise in the specified axis. Currently only axis=1 is supported.
- Parameters
- lowerscalar or array_like, default None
Minimum threshold value. All values below this threshold will be set to it. If it is None, there will be no clipping based on lower. In case of Series/Index, lower is expected to be a scalar or an array of size 1.
- upperscalar or array_like, default None
Maximum threshold value. All values below this threshold will be set to it. If it is None, there will be no clipping based on upper. In case of Series, upper is expected to be a scalar or an array of size 1.
- inplacebool, default False
- Returns
- Clipped DataFrame/Series/Index/MultiIndex
Examples
>>> import cudf >>> df = cudf.DataFrame({"a":[1, 2, 3, 4], "b":['a', 'b', 'c', 'd']}) >>> df.clip(lower=[2, 'b'], upper=[3, 'c']) a b 0 2 b 1 2 b 2 3 c 3 3 c
>>> df.clip(lower=None, upper=[3, 'c']) a b 0 1 a 1 2 b 2 3 c 3 3 c
>>> df.clip(lower=[2, 'b'], upper=None) a b 0 2 b 1 2 b 2 3 c 3 4 d
>>> df.clip(lower=2, upper=3, inplace=True) >>> df a b 0 2 2 1 2 3 2 3 3 3 3 3
>>> import cudf >>> sr = cudf.Series([1, 2, 3, 4]) >>> sr.clip(lower=2, upper=3) 0 2 1 2 2 3 3 3 dtype: int64
>>> sr.clip(lower=None, upper=3) 0 1 1 2 2 3 3 3 dtype: int64
>>> sr.clip(lower=2, upper=None, inplace=True) >>> sr 0 2 1 2 2 3 3 4 dtype: int64
-
property
columns¶ Returns a tuple of columns
-
copy(deep: bool = True) → T¶ Make a copy of this object’s indices and data.
When
deep=True(default), a new object will be created with a copy of the calling object’s data and indices. Modifications to the data or indices of the copy will not be reflected in the original object (see notes below). Whendeep=False, a new object will be created without copying the calling object’s data or index (only references to the data and index are copied). Any changes to the data of the original will be reflected in the shallow copy (and vice versa).- Parameters
- deepbool, default True
Make a deep copy, including a copy of the data and the indices. With
deep=Falseneither the indices nor the data are copied.
- Returns
- copySeries or DataFrame
Object type matches caller.
- Examples
>>> s = cudf.Series([1, 2], index=["a", "b"]) ..
>>> s ..
- a 1
- b 2
- dtype: int64
>>> s_copy = s.copy() ..
>>> s_copy ..
- a 1
- b 2
- dtype: int64
- Shallow copy versus default (deep) copy:
>>> s = cudf.Series([1, 2], index=["a", "b"]) ..
>>> deep = s.copy() ..
>>> shallow = s.copy(deep=False) ..
- Shallow copy shares data and index with original.
>>> s is shallow ..
- False
>>> s._column is shallow._column and s.index is shallow.index ..
- True
- Deep copy has own copy of data and index.
>>> s is deep ..
- False
>>> s.values is deep.values or s.index is deep.index ..
- False
- Updates to the data shared by shallow copy and original is reflected
- in both; deep copy remains unchanged.
>>> s['a'] = 3 ..
>>> shallow['b'] = 4 ..
>>> s ..
- a 3
- b 4
- dtype: int64
>>> shallow ..
- a 3
- b 4
- dtype: int64
>>> deep ..
- a 1
- b 2
- dtype: int64
-
corr()¶ Compute the correlation matrix of a DataFrame.
-
cos()¶ Get Trigonometric cosine, element-wise.
- Returns
- DataFrame/Series/Index
Result of the trigonometric operation.
Examples
>>> import cudf >>> ser = cudf.Series([0.0, 0.32434, 0.5, 45, 90, 180, 360]) >>> ser 0 0.00000 1 0.32434 2 0.50000 3 45.00000 4 90.00000 5 180.00000 6 360.00000 dtype: float64 >>> ser.cos() 0 1.000000 1 0.947861 2 0.877583 3 0.525322 4 -0.448074 5 -0.598460 6 -0.283691 dtype: float64
cos operation on DataFrame:
>>> df = cudf.DataFrame({'first': [0.0, 5, 10, 15], ... 'second': [100.0, 360, 720, 300]}) >>> df first second 0 0.0 100.0 1 5.0 360.0 2 10.0 720.0 3 15.0 300.0 >>> df.cos() first second 0 1.000000 0.862319 1 0.283662 -0.283691 2 -0.839072 -0.839039 3 -0.759688 -0.022097
cos operation on Index:
>>> index = cudf.Index([-0.4, 100, -180, 90]) >>> index Float64Index([-0.4, 100.0, -180.0, 90.0], dtype='float64') >>> index.cos() Float64Index([ 0.9210609940028851, 0.8623188722876839, -0.5984600690578581, -0.4480736161291701], dtype='float64')
-
count(axis=0, level=None, numeric_only=False, **kwargs)¶ Count
non-NAcells for each column or row.The values
None,NaN,NaTare consideredNA.- Returns
- Series
For each column/row the number of non-NA/null entries.
Notes
Parameters currently not supported are axis, level, numeric_only.
Examples
>>> import cudf >>> import numpy as np >>> df = cudf.DataFrame({"Person": ... ["John", "Myla", "Lewis", "John", "Myla"], ... "Age": [24., np.nan, 21., 33, 26], ... "Single": [False, True, True, True, False]}) >>> df.count() Person 5 Age 4 Single 5 dtype: int64
-
cov(**kwargs)¶ Compute the covariance matrix of a DataFrame.
- Parameters
- **kwargs
Keyword arguments to be passed to cupy.cov
- Returns
- covDataFrame
-
cummax(axis=None, skipna=True, *args, **kwargs)¶ Return cumulative maximum of the DataFrame.
- Parameters
- skipna: bool, default True
Exclude NA/null values. If an entire row/column is NA, the result will be NA.
- Returns
- DataFrame
Notes
Parameters currently not supported is axis
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.cummax() a b 0 1 7 1 2 8 2 3 9 3 4 10
-
cummin(axis=None, skipna=True, *args, **kwargs)¶ Return cumulative minimum of the DataFrame.
- Parameters
- skipna: bool, default True
Exclude NA/null values. If an entire row/column is NA, the result will be NA.
- Returns
- DataFrame
Notes
Parameters currently not supported is axis
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.cummin() a b 0 1 7 1 1 7 2 1 7 3 1 7
-
cumprod(axis=None, skipna=True, *args, **kwargs)¶ Return cumulative product of the DataFrame.
- Parameters
- skipna: bool, default True
Exclude NA/null values. If an entire row/column is NA, the result will be NA.
- Returns
- DataFrame
Notes
Parameters currently not supported is axis
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> s.cumprod() a b 0 1 7 1 2 56 2 6 504 3 24 5040
-
cumsum(axis=None, skipna=True, *args, **kwargs)¶ Return cumulative sum of the DataFrame.
- Parameters
- skipna: bool, default True
Exclude NA/null values. If an entire row/column is NA, the result will be NA.
- Returns
- DataFrame
Notes
Parameters currently not supported is axis
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> s.cumsum() a b 0 1 7 1 3 15 2 6 24 3 10 34
-
describe(percentiles=None, include=None, exclude=None, datetime_is_numeric=False)¶ Generate descriptive statistics.
Descriptive statistics include those that summarize the central tendency, dispersion and shape of a dataset’s distribution, excluding
NaNvalues.Analyzes both numeric and object series, as well as
DataFramecolumn sets of mixed data types. The output will vary depending on what is provided. Refer to the notes below for more detail.- Parameters
- percentileslist-like of numbers, optional
The percentiles to include in the output. All should fall between 0 and 1. The default is
[.25, .5, .75], which returns the 25th, 50th, and 75th percentiles.- include‘all’, list-like of dtypes or None(default), optional
A list of data types to include in the result. Ignored for
Series. Here are the options:‘all’ : All columns of the input will be included in the output.
A list-like of dtypes : Limits the results to the provided data types. To limit the result to numeric types submit
numpy.number. To limit it instead to object columns submit thenumpy.objectdata type. Strings can also be used in the style ofselect_dtypes(e.g.df.describe(include=['O'])). To select pandas categorical columns, use'category'None (default) : The result will include all numeric columns.
- excludelist-like of dtypes or None (default), optional,
A list of data types to omit from the result. Ignored for
Series. Here are the options:A list-like of dtypes : Excludes the provided data types from the result. To exclude numeric types submit
numpy.number. To exclude object columns submit the data typenumpy.object. Strings can also be used in the style ofselect_dtypes(e.g.df.describe(include=['O'])). To exclude pandas categorical columns, use'category'None (default) : The result will exclude nothing.
- datetime_is_numericbool, default False
For DataFrame input, this also controls whether datetime columns are included by default.
- Returns
- output_frameSeries or DataFrame
Summary statistics of the Series or Dataframe provided.
Notes
For numeric data, the result’s index will include
count,mean,std,min,maxas well as lower,50and upper percentiles. By default the lower percentile is25and the upper percentile is75. The50percentile is the same as the median.For strings dtype or datetime dtype, the result’s index will include
count,unique,top, andfreq. Thetopis the most common value. Thefreqis the most common value’s frequency. Timestamps also include thefirstandlastitems.If multiple object values have the highest count, then the
countandtopresults will be arbitrarily chosen from among those with the highest count.For mixed data types provided via a
DataFrame, the default is to return only an analysis of numeric columns. If the dataframe consists only of object and categorical data without any numeric columns, the default is to return an analysis of both the object and categorical columns. Ifinclude='all'is provided as an option, the result will include a union of attributes of each type.The
includeandexcludeparameters can be used to limit which columns in aDataFrameare analyzed for the output. The parameters are ignored when analyzing aSeries.Examples
Describing a
Seriescontaining numeric values.>>> import cudf >>> s = cudf.Series([1, 2, 3, 4, 5, 6, 7, 8, 9, 10]) >>> s 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 dtype: int64 >>> s.describe() count 10.00000 mean 5.50000 std 3.02765 min 1.00000 25% 3.25000 50% 5.50000 75% 7.75000 max 10.00000 dtype: float64
Describing a categorical
Series.>>> s = cudf.Series(['a', 'b', 'a', 'b', 'c', 'a'], dtype='category') >>> s 0 a 1 b 2 a 3 b 4 c 5 a dtype: category Categories (3, object): ['a', 'b', 'c'] >>> s.describe() count 6 unique 3 top a freq 3 dtype: object
Describing a timestamp
Series.>>> import numpy as np >>> s = cudf.Series([ ... np.datetime64("2000-01-01"), ... np.datetime64("2010-01-01"), ... np.datetime64("2010-01-01") ... ]) >>> s 0 2000-01-01 1 2010-01-01 2 2010-01-01 dtype: datetime64[s] >>> s.describe() count 3 mean 2006-09-01 08:00:00.000000000 min 2000-01-01 00:00:00.000000000 25% 2004-12-31 12:00:00.000000000 50% 2010-01-01 00:00:00.000000000 75% 2010-01-01 00:00:00.000000000 max 2010-01-01 00:00:00.000000000 dtype: object
Describing a
DataFrame. By default only numeric fields are returned.>>> df = cudf.DataFrame({"categorical": cudf.Series(['d', 'e', 'f'], ... dtype='category'), ... "numeric": [1, 2, 3], ... "object": ['a', 'b', 'c'] ... }) >>> df categorical numeric object 0 d 1 a 1 e 2 b 2 f 3 c >>> df.describe() numeric count 3.0 mean 2.0 std 1.0 min 1.0 25% 1.5 50% 2.0 75% 2.5 max 3.0
Describing all columns of a
DataFrameregardless of data type.>>> df.describe(include='all') categorical numeric object count 3 3.0 3 unique 3 <NA> 3 top d <NA> a freq 1 <NA> 1 mean <NA> 2.0 <NA> std <NA> 1.0 <NA> min <NA> 1.0 <NA> 25% <NA> 1.5 <NA> 50% <NA> 2.0 <NA> 75% <NA> 2.5 <NA> max <NA> 3.0 <NA>
Describing a column from a
DataFrameby accessing it as an attribute.>>> df.numeric.describe() count 3.0 mean 2.0 std 1.0 min 1.0 25% 1.5 50% 2.0 75% 2.5 max 3.0 Name: numeric, dtype: float64
Including only numeric columns in a
DataFramedescription.>>> df.describe(include=[np.number]) numeric count 3.0 mean 2.0 std 1.0 min 1.0 25% 1.5 50% 2.0 75% 2.5 max 3.0
Including only string columns in a
DataFramedescription.>>> df.describe(include=[object]) object count 3 unique 3 top a freq 1
Including only categorical columns from a
DataFramedescription.>>> df.describe(include=['category']) categorical count 3 unique 3 top d freq 1
Excluding numeric columns from a
DataFramedescription.>>> df.describe(exclude=[np.number]) categorical object count 3 3 unique 3 3 top d a freq 1 1
Excluding object columns from a
DataFramedescription.>>> df.describe(exclude=[object]) categorical numeric count 3 3.0 unique 3 <NA> top d <NA> freq 1 <NA> mean <NA> 2.0 std <NA> 1.0 min <NA> 1.0 25% <NA> 1.5 50% <NA> 2.0 75% <NA> 2.5 max <NA> 3.0
-
div(other, axis='columns', level=None, fill_value=None)¶ Get Floating division of dataframe and other, element-wise (binary operator truediv).
Equivalent to
dataframe / other, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, rtruediv.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df.truediv(10) angles degrees circle 0.0 36.0 triangle 0.3 18.0 rectangle 0.4 36.0 >>> df.div(10) angles degrees circle 0.0 36.0 triangle 0.3 18.0 rectangle 0.4 36.0 >>> df / 10 angles degrees circle 0.0 36.0 triangle 0.3 18.0 rectangle 0.4 36.0
-
drop(labels=None, axis=0, index=None, columns=None, level=None, inplace=False, errors='raise')¶ Drop specified labels from rows or columns.
Remove rows or columns by specifying label names and corresponding axis, or by specifying directly index or column names. When using a multi-index, labels on different levels can be removed by specifying the level.
- Parameters
- labelssingle label or list-like
Index or column labels to drop.
- axis{0 or ‘index’, 1 or ‘columns’}, default 0
Whether to drop labels from the index (0 or ‘index’) or columns (1 or ‘columns’).
- indexsingle label or list-like
Alternative to specifying axis (
labels, axis=0is equivalent toindex=labels).- columnssingle label or list-like
Alternative to specifying axis (
labels, axis=1is equivalent tocolumns=labels).- levelint or level name, optional
For MultiIndex, level from which the labels will be removed.
- inplacebool, default False
If False, return a copy. Otherwise, do operation inplace and return None.
- errors{‘ignore’, ‘raise’}, default ‘raise’
If ‘ignore’, suppress error and only existing labels are dropped.
- Returns
- DataFrame
DataFrame without the removed index or column labels.
- Raises
- KeyError
If any of the labels is not found in the selected axis.
See also
DataFrame.locLabel-location based indexer for selection by label.
DataFrame.dropnaReturn DataFrame with labels on given axis omitted where (all or any) data are missing.
DataFrame.drop_duplicatesReturn DataFrame with duplicate rows removed, optionally only considering certain columns.
Examples
>>> import cudf >>> df = cudf.DataFrame({"A": [1, 2, 3, 4], ... "B": [5, 6, 7, 8], ... "C": [10, 11, 12, 13], ... "D": [20, 30, 40, 50]}) >>> df A B C D 0 1 5 10 20 1 2 6 11 30 2 3 7 12 40 3 4 8 13 50
Drop columns
>>> df.drop(['B', 'C'], axis=1) A D 0 1 20 1 2 30 2 3 40 3 4 50 >>> df.drop(columns=['B', 'C']) A D 0 1 20 1 2 30 2 3 40 3 4 50
Drop a row by index
>>> df.drop([0, 1]) A B C D 2 3 7 12 40 3 4 8 13 50
Drop columns and/or rows of MultiIndex DataFrame
>>> midx = cudf.MultiIndex(levels=[['lama', 'cow', 'falcon'], ... ['speed', 'weight', 'length']], ... codes=[[0, 0, 0, 1, 1, 1, 2, 2, 2], ... [0, 1, 2, 0, 1, 2, 0, 1, 2]]) >>> df = cudf.DataFrame(index=midx, columns=['big', 'small'], ... data=[[45, 30], [200, 100], [1.5, 1], [30, 20], ... [250, 150], [1.5, 0.8], [320, 250], ... [1, 0.8], [0.3, 0.2]]) >>> df big small lama speed 45.0 30.0 weight 200.0 100.0 length 1.5 1.0 cow speed 30.0 20.0 weight 250.0 150.0 length 1.5 0.8 falcon speed 320.0 250.0 weight 1.0 0.8 length 0.3 0.2 >>> df.drop(index='cow', columns='small') big lama speed 45.0 weight 200.0 length 1.5 falcon speed 320.0 weight 1.0 length 0.3 >>> df.drop(index='length', level=1) big small lama speed 45.0 30.0 weight 200.0 100.0 cow speed 30.0 20.0 weight 250.0 150.0 falcon speed 320.0 250.0 weight 1.0 0.8
-
drop_duplicates(subset=None, keep='first', inplace=False, ignore_index=False)¶ Return DataFrame with duplicate rows removed, optionally only considering certain subset of columns.
-
dropna(axis=0, how='any', thresh=None, subset=None, inplace=False)¶ Drops rows (or columns) containing nulls from a Column.
- Parameters
- axis{0, 1}, optional
Whether to drop rows (axis=0, default) or columns (axis=1) containing nulls.
- how{“any”, “all”}, optional
Specifies how to decide whether to drop a row (or column). any (default) drops rows (or columns) containing at least one null value. all drops only rows (or columns) containing all null values.
- thresh: int, optional
If specified, then drops every row (or column) containing less than thresh non-null values
- subsetlist, optional
List of columns to consider when dropping rows (all columns are considered by default). Alternatively, when dropping columns, subset is a list of rows to consider.
- inplacebool, default False
If True, do operation inplace and return None.
- Returns
- Copy of the DataFrame with rows/columns containing nulls dropped.
See also
cudf.core.dataframe.DataFrame.isnaIndicate null values.
cudf.core.dataframe.DataFrame.notnaIndicate non-null values.
cudf.core.dataframe.DataFrame.fillnaReplace null values.
cudf.core.series.Series.dropnaDrop null values.
cudf.core.index.Index.dropnaDrop null indices.
Examples
>>> import cudf >>> df = cudf.DataFrame({"name": ['Alfred', 'Batman', 'Catwoman'], ... "toy": ['Batmobile', None, 'Bullwhip'], ... "born": [np.datetime64("1940-04-25"), ... np.datetime64("NaT"), ... np.datetime64("NaT")]}) >>> df name toy born 0 Alfred Batmobile 1940-04-25 00:00:00 1 Batman <NA> <NA> 2 Catwoman Bullwhip <NA>
Drop the rows where at least one element is null.
>>> df.dropna() name toy born 0 Alfred Batmobile 1940-04-25
Drop the columns where at least one element is null.
>>> df.dropna(axis='columns') name 0 Alfred 1 Batman 2 Catwoman
Drop the rows where all elements are null.
>>> df.dropna(how='all') name toy born 0 Alfred Batmobile 1940-04-25 00:00:00 1 Batman <NA> <NA> 2 Catwoman Bullwhip <NA>
Keep only the rows with at least 2 non-null values.
>>> df.dropna(thresh=2) name toy born 0 Alfred Batmobile 1940-04-25 00:00:00 2 Catwoman Bullwhip <NA>
Define in which columns to look for null values.
>>> df.dropna(subset=['name', 'born']) name toy born 0 Alfred Batmobile 1940-04-25
Keep the DataFrame with valid entries in the same variable.
>>> df.dropna(inplace=True) >>> df name toy born 0 Alfred Batmobile 1940-04-25
-
property
dtypes¶ Return the dtypes in this object.
-
property
empty¶ Indicator whether DataFrame or Series is empty.
True if DataFrame/Series is entirely empty (no items), meaning any of the axes are of length 0.
- Returns
- outbool
If DataFrame/Series is empty, return True, if not return False.
Notes
If DataFrame/Series contains only null values, it is still not considered empty. See the example below.
Examples
>>> import cudf >>> df = cudf.DataFrame({'A' : []}) >>> df Empty DataFrame Columns: [A] Index: [] >>> df.empty True
If we only have null values in our DataFrame, it is not considered empty! We will need to drop the null’s to make the DataFrame empty:
>>> df = cudf.DataFrame({'A' : [None, None]}) >>> df A 0 <NA> 1 <NA> >>> df.empty False >>> df.dropna().empty True
Non-empty and empty Series example:
>>> s = cudf.Series([1, 2, None]) >>> s 0 1 1 2 2 <NA> dtype: int64 >>> s.empty False >>> s = cudf.Series([]) >>> s Series([], dtype: float64) >>> s.empty True
-
equals(other)¶ Test whether two objects contain the same elements. This function allows two Series or DataFrames to be compared against each other to see if they have the same shape and elements. NaNs in the same location are considered equal. The column headers do not need to have the same type.
- Parameters
- otherSeries or DataFrame
The other Series or DataFrame to be compared with the first.
- Returns
- bool
True if all elements are the same in both objects, False otherwise.
Examples
>>> import cudf
Comparing Series with equals:
>>> s = cudf.Series([1, 2, 3]) >>> other = cudf.Series([1, 2, 3]) >>> s.equals(other) True >>> different = cudf.Series([1.5, 2, 3]) >>> s.equals(different) False
Comparing DataFrames with equals:
>>> df = cudf.DataFrame({1: [10], 2: [20]}) >>> df 1 2 0 10 20 >>> exactly_equal = cudf.DataFrame({1: [10], 2: [20]}) >>> exactly_equal 1 2 0 10 20 >>> df.equals(exactly_equal) True
For two DataFrames to compare equal, the types of column values must be equal, but the types of column labels need not:
>>> different_column_type = cudf.DataFrame({1.0: [10], 2.0: [20]}) >>> different_column_type 1.0 2.0 0 10 20 >>> df.equals(different_column_type) True
-
exp()¶ Get the exponential of all elements, element-wise.
Exponential is the inverse of the log function, so that x.exp().log() = x
- Returns
- DataFrame/Series/Index
Result of the element-wise exponential.
Examples
>>> import cudf >>> ser = cudf.Series([-1, 0, 1, 0.32434, 0.5, -10, 100]) >>> ser 0 -1.00000 1 0.00000 2 1.00000 3 0.32434 4 0.50000 5 -10.00000 6 100.00000 dtype: float64 >>> ser.exp() 0 3.678794e-01 1 1.000000e+00 2 2.718282e+00 3 1.383117e+00 4 1.648721e+00 5 4.539993e-05 6 2.688117e+43 dtype: float64
exp operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-1, -10, 0.5], ... 'second': [0.234, 0.3, 10]}) >>> df first second 0 -1.0 0.234 1 -10.0 0.300 2 0.5 10.000 >>> df.exp() first second 0 0.367879 1.263644 1 0.000045 1.349859 2 1.648721 22026.465795
exp operation on Index:
>>> index = cudf.Index([-1, 0.4, 1, 0, 0.3]) >>> index Float64Index([-1.0, 0.4, 1.0, 0.0, 0.3], dtype='float64') >>> index.exp() Float64Index([0.36787944117144233, 1.4918246976412703, 2.718281828459045, 1.0, 1.3498588075760032], dtype='float64')
-
fillna(value=None, method=None, axis=None, inplace=False, limit=None)¶ Fill null values with
valueor specifiedmethod.- Parameters
- valuescalar, Series-like or dict
Value to use to fill nulls. If Series-like, null values are filled with values in corresponding indices. A dict can be used to provide different values to fill nulls in different columns. Cannot be used with
method.- method{‘ffill’, ‘bfill’}, default None
Method to use for filling null values in the dataframe or series. ffill propagates the last non-null values forward to the next non-null value. bfill propagates backward with the next non-null value. Cannot be used with
value.
- Returns
- resultDataFrame
Copy with nulls filled.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, None], 'b': [3, None, 5]}) >>> df a b 0 1 3 1 2 <NA> 2 <NA> 5 >>> df.fillna(4) a b 0 1 3 1 2 4 2 4 5 >>> df.fillna({'a': 3, 'b': 4}) a b 0 1 3 1 2 4 2 3 5
fillnaon a Series object:>>> ser = cudf.Series(['a', 'b', None, 'c']) >>> ser 0 a 1 b 2 <NA> 3 c dtype: object >>> ser.fillna('z') 0 a 1 b 2 z 3 c dtype: object
fillnacan also supports inplace operation:>>> ser.fillna('z', inplace=True) >>> ser 0 a 1 b 2 z 3 c dtype: object >>> df.fillna({'a': 3, 'b': 4}, inplace=True) >>> df a b 0 1 3 1 2 4 2 3 5
fillnaspecified with fillmethod>>> ser = cudf.Series([1, None, None, 2, 3, None, None]) >>> ser.fillna(method='ffill') 0 1 1 1 2 1 3 2 4 3 5 3 6 3 dtype: int64 >>> ser.fillna(method='bfill') 0 1 1 2 2 2 3 2 4 3 5 <NA> 6 <NA> dtype: int64
-
floordiv(other, axis='columns', level=None, fill_value=None)¶ Get Integer division of dataframe and other, element-wise (binary operator floordiv).
Equivalent to
dataframe // other, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, rfloordiv.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [1, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df.floordiv(2) angles degrees circle 0 180 triangle 1 90 rectangle 2 180 >>> df // 2 angles degrees circle 0 180 triangle 1 90 rectangle 2 180
-
classmethod
from_arrow(table)¶ Convert from PyArrow Table to DataFrame.
- Parameters
- tablePyArrow Table Object
PyArrow Table Object which has to be converted to cudf DataFrame.
- Returns
- cudf DataFrame
- Raises
- TypeError for invalid input type.
Notes
Does not support automatically setting index column(s) similar to how
to_pandasworks for PyArrow Tables.
Examples
>>> import cudf >>> import pyarrow as pa >>> data = pa.table({"a":[1, 2, 3], "b":[4, 5, 6]}) >>> cudf.DataFrame.from_arrow(data) a b 0 1 4 1 2 5 2 3 6
-
classmethod
from_pandas(dataframe, nan_as_null=None)¶ Convert from a Pandas DataFrame.
- Parameters
- dataframePandas DataFrame object
A Pandads DataFrame object which has to be converted to cuDF DataFrame.
- nan_as_nullbool, Default True
If
True, convertsnp.nanvalues tonullvalues. IfFalse, leavesnp.nanvalues as is.
- Raises
- TypeError for invalid input type.
Examples
>>> import cudf >>> import pandas as pd >>> data = [[0,1], [1,2], [3,4]] >>> pdf = pd.DataFrame(data, columns=['a', 'b'], dtype=int) >>> cudf.from_pandas(pdf) a b 0 0 1 1 1 2 2 3 4
-
classmethod
from_records(data, index=None, columns=None, nan_as_null=False)¶ Convert structured or record ndarray to DataFrame.
- Parameters
- datanumpy structured dtype or recarray of ndim=2
- indexstr, array-like
The name of the index column in data. If None, the default index is used.
- columnslist of str
List of column names to include.
- Returns
- DataFrame
-
hash_columns(columns=None)¶ Hash the given columns and return a new device array
- Parameters
- columnssequence of str; optional
Sequence of column names. If columns is None (unspecified), all columns in the frame are used.
-
head(n=5)¶ Returns the first n rows as a new DataFrame
Examples
>>> import cudf >>> df = cudf.DataFrame() >>> df['key'] = [0, 1, 2, 3, 4] >>> df['val'] = [float(i + 10) for i in range(5)] # insert column >>> df.head(2) key val 0 0 10.0 1 1 11.0
-
property
iat¶ Alias for
DataFrame.iloc; provided for compatibility with Pandas.
-
property
iloc¶ Selecting rows and column by position.
See also
Notes
One notable difference from Pandas is when DataFrame is of mixed types and result is expected to be a Series in case of Pandas. cuDF will return a DataFrame as it doesn’t support mixed types under Series yet.
Mixed dtype single row output as a dataframe (pandas results in Series)
>>> import cudf >>> df = cudf.DataFrame({"a":[1, 2, 3], "b":["a", "b", "c"]}) >>> df.iloc[0] a b 0 1 a
Examples
>>> df = cudf.DataFrame({'a': range(20), ... 'b': range(20), ... 'c': range(20)})
Select a single row using an integer index.
>>> df.iloc[1] a 1 b 1 c 1 Name: 1, dtype: int64
Select multiple rows using a list of integers.
>>> df.iloc[[0, 2, 9, 18]] a b c 0 0 0 0 2 2 2 2 9 9 9 9 18 18 18 18
Select rows using a slice.
>>> df.iloc[3:10:2] a b c 3 3 3 3 5 5 5 5 7 7 7 7 9 9 9 9
Select both rows and columns.
>>> df.iloc[[1, 3, 5, 7], 2] 1 1 3 3 5 5 7 7 Name: c, dtype: int64
Setting values in a column using iloc.
>>> df.iloc[:4] = 0 >>> df a b c 0 0 0 0 1 0 0 0 2 0 0 0 3 0 0 0 4 4 4 4 5 5 5 5 6 6 6 6 7 7 7 7 8 8 8 8 9 9 9 9 [10 more rows]
-
property
index¶ Returns the index of the DataFrame
-
info(verbose=None, buf=None, max_cols=None, memory_usage=None, null_counts=None)¶ Print a concise summary of a DataFrame.
This method prints information about a DataFrame including the index dtype and column dtypes, non-null values and memory usage.
- Parameters
- verbosebool, optional
Whether to print the full summary. By default, the setting in
pandas.options.display.max_info_columnsis followed.- bufwritable buffer, defaults to sys.stdout
Where to send the output. By default, the output is printed to sys.stdout. Pass a writable buffer if you need to further process the output.
- max_colsint, optional
When to switch from the verbose to the truncated output. If the DataFrame has more than max_cols columns, the truncated output is used. By default, the setting in
pandas.options.display.max_info_columnsis used.- memory_usagebool, str, optional
Specifies whether total memory usage of the DataFrame elements (including the index) should be displayed. By default, this follows the
pandas.options.display.memory_usagesetting. True always show memory usage. False never shows memory usage. A value of ‘deep’ is equivalent to “True with deep introspection”. Memory usage is shown in human-readable units (base-2 representation). Without deep introspection a memory estimation is made based in column dtype and number of rows assuming values consume the same memory amount for corresponding dtypes. With deep memory introspection, a real memory usage calculation is performed at the cost of computational resources.- null_countsbool, optional
Whether to show the non-null counts. By default, this is shown only if the frame is smaller than
pandas.options.display.max_info_rowsandpandas.options.display.max_info_columns. A value of True always shows the counts, and False never shows the counts.
- Returns
- None
This method prints a summary of a DataFrame and returns None.
See also
DataFrame.describeGenerate descriptive statistics of DataFrame columns.
DataFrame.memory_usageMemory usage of DataFrame columns.
Examples
>>> import cudf >>> int_values = [1, 2, 3, 4, 5] >>> text_values = ['alpha', 'beta', 'gamma', 'delta', 'epsilon'] >>> float_values = [0.0, 0.25, 0.5, 0.75, 1.0] >>> df = cudf.DataFrame({"int_col": int_values, ... "text_col": text_values, ... "float_col": float_values}) >>> df int_col text_col float_col 0 1 alpha 0.00 1 2 beta 0.25 2 3 gamma 0.50 3 4 delta 0.75 4 5 epsilon 1.00
Prints information of all columns:
>>> df.info(verbose=True) <class 'cudf.core.dataframe.DataFrame'> RangeIndex: 5 entries, 0 to 4 Data columns (total 3 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 int_col 5 non-null int64 1 text_col 5 non-null object 2 float_col 5 non-null float64 dtypes: float64(1), int64(1), object(1) memory usage: 130.0+ bytes
Prints a summary of columns count and its dtypes but not per column information:
>>> df.info(verbose=False) <class 'cudf.core.dataframe.DataFrame'> RangeIndex: 5 entries, 0 to 4 Columns: 3 entries, int_col to float_col dtypes: float64(1), int64(1), object(1) memory usage: 130.0+ bytes
Pipe output of DataFrame.info to buffer instead of sys.stdout, get buffer content and writes to a text file:
>>> import io >>> buffer = io.StringIO() >>> df.info(buf=buffer) >>> s = buffer.getvalue() >>> with open("df_info.txt", "w", ... encoding="utf-8") as f: ... f.write(s) ... 369
The memory_usage parameter allows deep introspection mode, specially useful for big DataFrames and fine-tune memory optimization:
>>> import numpy as np >>> random_strings_array = np.random.choice(['a', 'b', 'c'], 10 ** 6) >>> df = cudf.DataFrame({ ... 'column_1': np.random.choice(['a', 'b', 'c'], 10 ** 6), ... 'column_2': np.random.choice(['a', 'b', 'c'], 10 ** 6), ... 'column_3': np.random.choice(['a', 'b', 'c'], 10 ** 6) ... }) >>> df.info(memory_usage='deep') <class 'cudf.core.dataframe.DataFrame'> RangeIndex: 1000000 entries, 0 to 999999 Data columns (total 3 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 column_1 1000000 non-null object 1 column_2 1000000 non-null object 2 column_3 1000000 non-null object dtypes: object(3) memory usage: 14.3 MB
-
insert(loc, name, value)¶ Add a column to DataFrame at the index specified by loc.
- Parameters
- locint
location to insert by index, cannot be greater then num columns + 1
- namenumber or string
name or label of column to be inserted
- valueSeries or array-like
-
interleave_columns()¶ Interleave Series columns of a table into a single column.
Converts the column major table cols into a row major column.
- Parameters
- colsinput Table containing columns to interleave.
- Returns
- The interleaved columns as a single column
Examples
>>> df = DataFrame([['A1', 'A2', 'A3'], ['B1', 'B2', 'B3']]) >>> df 0 [A1, A2, A3] 1 [B1, B2, B3] >>> df.interleave_columns() 0 A1 1 B1 2 A2 3 B2 4 A3 5 B3
-
isin(values)¶ Whether each element in the DataFrame is contained in values.
- Parameters
- valuesiterable, Series, DataFrame or dict
The result will only be true at a location if all the labels match. If values is a Series, that’s the index. If values is a dict, the keys must be the column names, which must match. If values is a DataFrame, then both the index and column labels must match.
- Returns
- DataFrame:
DataFrame of booleans showing whether each element in the DataFrame is contained in values.
-
isna()¶ Identify missing values.
Return a boolean same-sized object indicating if the values are
<NA>.<NA>values gets mapped toTruevalues. Everything else gets mapped toFalsevalues.<NA>values include:Values where null mask is set.
NaNin float dtype.NaTin datetime64 and timedelta64 types.
Characters such as empty strings
''orinfincase of float are not considered<NA>values.- Returns
- DataFrame/Series/Index
Mask of bool values for each element in the object that indicates whether an element is an NA value.
Examples
Show which entries in a DataFrame are NA.
>>> import cudf >>> import numpy as np >>> import pandas as pd >>> df = cudf.DataFrame({'age': [5, 6, np.NaN], ... 'born': [pd.NaT, pd.Timestamp('1939-05-27'), ... pd.Timestamp('1940-04-25')], ... 'name': ['Alfred', 'Batman', ''], ... 'toy': [None, 'Batmobile', 'Joker']}) >>> df age born name toy 0 5 <NA> Alfred <NA> 1 6 1939-05-27 00:00:00.000000 Batman Batmobile 2 <NA> 1940-04-25 00:00:00.000000 Joker >>> df.isnull() age born name toy 0 False True False True 1 False False False False 2 True False False False
Show which entries in a Series are NA.
>>> ser = cudf.Series([5, 6, np.NaN, np.inf, -np.inf]) >>> ser 0 5.0 1 6.0 2 <NA> 3 Inf 4 -Inf dtype: float64 >>> ser.isnull() 0 False 1 False 2 True 3 False 4 False dtype: bool
Show which entries in an Index are NA.
>>> idx = cudf.Index([1, 2, None, np.NaN, 0.32, np.inf]) >>> idx Float64Index([1.0, 2.0, <NA>, <NA>, 0.32, Inf], dtype='float64') >>> idx.isnull() GenericIndex([False, False, True, True, False, False], dtype='bool')
-
isnull()¶ Identify missing values.
Return a boolean same-sized object indicating if the values are
<NA>.<NA>values gets mapped toTruevalues. Everything else gets mapped toFalsevalues.<NA>values include:Values where null mask is set.
NaNin float dtype.NaTin datetime64 and timedelta64 types.
Characters such as empty strings
''orinfincase of float are not considered<NA>values.- Returns
- DataFrame/Series/Index
Mask of bool values for each element in the object that indicates whether an element is an NA value.
Examples
Show which entries in a DataFrame are NA.
>>> import cudf >>> import numpy as np >>> import pandas as pd >>> df = cudf.DataFrame({'age': [5, 6, np.NaN], ... 'born': [pd.NaT, pd.Timestamp('1939-05-27'), ... pd.Timestamp('1940-04-25')], ... 'name': ['Alfred', 'Batman', ''], ... 'toy': [None, 'Batmobile', 'Joker']}) >>> df age born name toy 0 5 <NA> Alfred <NA> 1 6 1939-05-27 00:00:00.000000 Batman Batmobile 2 <NA> 1940-04-25 00:00:00.000000 Joker >>> df.isnull() age born name toy 0 False True False True 1 False False False False 2 True False False False
Show which entries in a Series are NA.
>>> ser = cudf.Series([5, 6, np.NaN, np.inf, -np.inf]) >>> ser 0 5.0 1 6.0 2 <NA> 3 Inf 4 -Inf dtype: float64 >>> ser.isnull() 0 False 1 False 2 True 3 False 4 False dtype: bool
Show which entries in an Index are NA.
>>> idx = cudf.Index([1, 2, None, np.NaN, 0.32, np.inf]) >>> idx Float64Index([1.0, 2.0, <NA>, <NA>, 0.32, Inf], dtype='float64') >>> idx.isnull() GenericIndex([False, False, True, True, False, False], dtype='bool')
-
iteritems()¶ Iterate over column names and series pairs
-
join(other, on=None, how='left', lsuffix='', rsuffix='', sort=False, method='hash')¶ Join columns with other DataFrame on index or on a key column.
- Parameters
- otherDataFrame
- howstr
Only accepts “left”, “right”, “inner”, “outer”
- lsuffix, rsuffixstr
The suffices to add to the left (lsuffix) and right (rsuffix) column names when avoiding conflicts.
- sortbool
Set to True to ensure sorted ordering.
- Returns
- joinedDataFrame
Notes
Difference from pandas:
other must be a single DataFrame for now.
on is not supported yet due to lack of multi-index support.
-
keys()¶ Get the columns. This is index for Series, columns for DataFrame.
- Returns
- Index
Columns of DataFrame.
Examples
>>> import cudf >>> df = cudf.DataFrame({'one' : [1, 2, 3], 'five' : ['a', 'b', 'c']}) >>> df one five 0 1 a 1 2 b 2 3 c >>> df.keys() Index(['one', 'five'], dtype='object') >>> df = cudf.DataFrame(columns=[0, 1, 2, 3]) >>> df Empty DataFrame Columns: [0, 1, 2, 3] Index: [] >>> df.keys() Int64Index([0, 1, 2, 3], dtype='int64')
-
kurt(axis=None, skipna=None, level=None, numeric_only=None, **kwargs)¶ Return Fisher’s unbiased kurtosis of a sample.
Kurtosis obtained using Fisher’s definition of kurtosis (kurtosis of normal == 0.0). Normalized by N-1.
- Parameters
- skipna: bool, default True
Exclude NA/null values when computing the result.
- Returns
- Series
Notes
Parameters currently not supported are axis, level and numeric_only
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.kurt() a -1.2 b -1.2 dtype: float64
-
kurtosis(axis=None, skipna=None, level=None, numeric_only=None, **kwargs)¶ Return Fisher’s unbiased kurtosis of a sample.
Kurtosis obtained using Fisher’s definition of kurtosis (kurtosis of normal == 0.0). Normalized by N-1.
- Parameters
- skipna: bool, default True
Exclude NA/null values when computing the result.
- Returns
- Series
Notes
Parameters currently not supported are axis, level and numeric_only
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.kurt() a -1.2 b -1.2 dtype: float64
-
label_encoding(column, prefix, cats, prefix_sep='_', dtype=None, na_sentinel=- 1)¶ Encode labels in a column with label encoding.
- Parameters
- columnstr
the source column with binary encoding for the data.
- prefixstr
the new column name prefix.
- catssequence of ints
the sequence of categories as integers.
- prefix_sepstr
the separator between the prefix and the category.
- dtype :
the dtype for the outputs; see Series.label_encoding
- na_sentinelnumber
Value to indicate missing category.
- Returns
- a new dataframe with a new column append for the coded values.
-
property
loc¶ Selecting rows and columns by label or boolean mask.
See also
Notes
One notable difference from Pandas is when DataFrame is of mixed types and result is expected to be a Series in case of Pandas. cuDF will return a DataFrame as it doesn’t support mixed types under Series yet.
Mixed dtype single row output as a dataframe (pandas results in Series)
>>> import cudf >>> df = cudf.DataFrame({"a":[1, 2, 3], "b":["a", "b", "c"]}) >>> df.loc[0] a b 0 1 a
Examples
DataFrame with string index.
>>> df a b a 0 5 b 1 6 c 2 7 d 3 8 e 4 9
Select a single row by label.
>>> df.loc['a'] a 0 b 5 Name: a, dtype: int64
Select multiple rows and a single column.
>>> df.loc[['a', 'c', 'e'], 'b'] a 5 c 7 e 9 Name: b, dtype: int64
Selection by boolean mask.
>>> df.loc[df.a > 2] a b d 3 8 e 4 9
Setting values using loc.
>>> df.loc[['a', 'c', 'e'], 'a'] = 0 >>> df a b a 0 5 b 1 6 c 0 7 d 3 8 e 0 9
-
log()¶ Get the natural logarithm of all elements, element-wise.
Natural logarithm is the inverse of the exp function, so that x.log().exp() = x
- Returns
- DataFrame/Series/Index
Result of the element-wise natural logarithm.
Examples
>>> import cudf >>> ser = cudf.Series([-1, 0, 1, 0.32434, 0.5, -10, 100]) >>> ser 0 -1.00000 1 0.00000 2 1.00000 3 0.32434 4 0.50000 5 -10.00000 6 100.00000 dtype: float64 >>> ser.log() 0 NaN 1 -inf 2 0.000000 3 -1.125963 4 -0.693147 5 NaN 6 4.605170 dtype: float64
log operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-1, -10, 0.5], ... 'second': [0.234, 0.3, 10]}) >>> df first second 0 -1.0 0.234 1 -10.0 0.300 2 0.5 10.000 >>> df.log() first second 0 NaN -1.452434 1 NaN -1.203973 2 -0.693147 2.302585
log operation on Index:
>>> index = cudf.Index([10, 11, 500.0]) >>> index Float64Index([10.0, 11.0, 500.0], dtype='float64') >>> index.log() Float64Index([2.302585092994046, 2.3978952727983707, 6.214608098422191], dtype='float64')
-
mask(cond, other=None, inplace=False)¶ Replace values where the condition is True.
- Parameters
- condbool Series/DataFrame, array-like
Where cond is False, keep the original value. Where True, replace with corresponding value from other. Callables are not supported.
- other: scalar, list of scalars, Series/DataFrame
Entries where cond is True are replaced with corresponding value from other. Callables are not supported. Default is None.
DataFrame expects only Scalar or array like with scalars or dataframe with same dimension as self.
Series expects only scalar or series like with same length
- inplacebool, default False
Whether to perform the operation in place on the data.
- Returns
- Same type as caller
Examples
>>> import cudf >>> df = cudf.DataFrame({"A":[1, 4, 5], "B":[3, 5, 8]}) >>> df.mask(df % 2 == 0, [-1, -1]) A B 0 1 3 1 -1 5 2 5 -1
>>> ser = cudf.Series([4, 3, 2, 1, 0]) >>> ser.mask(ser > 2, 10) 0 10 1 10 2 2 3 1 4 0 dtype: int64 >>> ser.mask(ser > 2) 0 <NA> 1 <NA> 2 2 3 1 4 0 dtype: int64
-
max(axis=None, skipna=None, level=None, numeric_only=None, **kwargs)¶ Return the maximum of the values in the DataFrame.
- Parameters
- axis: {index (0), columns(1)}
Axis for the function to be applied on.
- skipna: bool, default True
Exclude NA/null values when computing the result.
- level: int or level name, default None
If the axis is a MultiIndex (hierarchical), count along a particular level, collapsing into a Series.
- numeric_only: bool, default None
Include only float, int, boolean columns. If None, will attempt to use everything, then use only numeric data.
- Returns
- Series
Notes
Parameters currently not supported are level, numeric_only.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.max() a 4 b 10 dtype: int64
-
mean(axis=None, skipna=None, level=None, numeric_only=None, **kwargs)¶ Return the mean of the values for the requested axis.
- Parameters
- axis{0 or ‘index’, 1 or ‘columns’}
Axis for the function to be applied on.
- skipnabool, default True
Exclude NA/null values when computing the result.
- levelint or level name, default None
If the axis is a MultiIndex (hierarchical), count along a particular level, collapsing into a Series.
- numeric_onlybool, default None
Include only float, int, boolean columns. If None, will attempt to use everything, then use only numeric data. Not implemented for Series.
- **kwargs
Additional keyword arguments to be passed to the function.
- Returns
- meanSeries or DataFrame (if level specified)
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.mean() a 2.5 b 8.5 dtype: float64
-
melt(**kwargs)¶ Unpivots a DataFrame from wide format to long format, optionally leaving identifier variables set.
- Parameters
- frameDataFrame
- id_varstuple, list, or ndarray, optional
Column(s) to use as identifier variables. default: None
- value_varstuple, list, or ndarray, optional
Column(s) to unpivot. default: all columns that are not set as id_vars.
- var_namescalar
Name to use for the variable column. default: frame.columns.name or ‘variable’
- value_namestr
Name to use for the value column. default: ‘value’
- Returns
- outDataFrame
Melted result
-
memory_usage(index=True, deep=False)¶ Return the memory usage of each column in bytes. The memory usage can optionally include the contribution of the index and elements of object dtype.
- Parameters
- indexbool, default True
Specifies whether to include the memory usage of the DataFrame’s index in returned Series. If
index=True, the memory usage of the index is the first item in the output.- deepbool, default False
If True, introspect the data deeply by interrogating object dtypes for system-level memory consumption, and include it in the returned values.
- Returns
- Series
A Series whose index is the original column names and whose values is the memory usage of each column in bytes.
Examples
>>> dtypes = ['int64', 'float64', 'object', 'bool'] >>> data = dict([(t, np.ones(shape=5000).astype(t)) ... for t in dtypes]) >>> df = cudf.DataFrame(data) >>> df.head() int64 float64 object bool 0 1 1.0 1.0 True 1 1 1.0 1.0 True 2 1 1.0 1.0 True 3 1 1.0 1.0 True 4 1 1.0 1.0 True >>> df.memory_usage(index=False) int64 40000 float64 40000 object 40000 bool 5000 dtype: int64 Use a Categorical for efficient storage of an object-dtype column with many repeated values. >>> df['object'].astype('category').memory_usage(deep=True) 5048
-
merge(right, on=None, left_on=None, right_on=None, left_index=False, right_index=False, how='inner', sort=False, lsuffix=None, rsuffix=None, method='hash', indicator=False, suffixes=('_x', '_y'))¶ Merge GPU DataFrame objects by performing a database-style join operation by columns or indexes.
- Parameters
- rightDataFrame
- onlabel or list; defaults to None
Column or index level names to join on. These must be found in both DataFrames.
If on is None and not merging on indexes then this defaults to the intersection of the columns in both DataFrames.
- how{‘left’, ‘outer’, ‘inner’}, default ‘inner’
Type of merge to be performed.
left : use only keys from left frame, similar to a SQL left outer join.
right : not supported.
outer : use union of keys from both frames, similar to a SQL full outer join.
inner: use intersection of keys from both frames, similar to a SQL inner join.
- left_onlabel or list, or array-like
Column or index level names to join on in the left DataFrame. Can also be an array or list of arrays of the length of the left DataFrame. These arrays are treated as if they are columns.
- right_onlabel or list, or array-like
Column or index level names to join on in the right DataFrame. Can also be an array or list of arrays of the length of the right DataFrame. These arrays are treated as if they are columns.
- left_indexbool, default False
Use the index from the left DataFrame as the join key(s).
- right_indexbool, default False
Use the index from the right DataFrame as the join key.
- sortbool, default False
Sort the resulting dataframe by the columns that were merged on, starting from the left.
- suffixes: Tuple[str, str], defaults to (‘_x’, ‘_y’)
Suffixes applied to overlapping column names on the left and right sides
- method{‘hash’, ‘sort’}, default ‘hash’
The implementation method to be used for the operation.
- Returns
- mergedDataFrame
Notes
DataFrames merges in cuDF result in non-deterministic row ordering.
Examples
>>> import cudf >>> df_a = cudf.DataFrame() >>> df_a['key'] = [0, 1, 2, 3, 4] >>> df_a['vals_a'] = [float(i + 10) for i in range(5)] >>> df_b = cudf.DataFrame() >>> df_b['key'] = [1, 2, 4] >>> df_b['vals_b'] = [float(i+10) for i in range(3)] >>> df_merged = df_a.merge(df_b, on=['key'], how='left') >>> df_merged.sort_values('key') key vals_a vals_b 3 0 10.0 0 1 11.0 10.0 1 2 12.0 11.0 4 3 13.0 2 4 14.0 12.0
Merging on categorical variables is only allowed in certain cases
Categorical variable typecasting logic depends on both how and the specifics of the categorical variables to be merged. Merging categorical variables when only one side is ordered is ambiguous and not allowed. Merging when both categoricals are ordered is allowed, but only when the categories are exactly equal and have equal ordering, and will result in the common dtype. When both sides are unordered, the result categorical depends on the kind of join: - For inner joins, the result will be the intersection of the categories - For left or right joins, the result will be the the left or right dtype respectively. This extends to semi and anti joins. - For outer joins, the result will be the union of categories from both sides.
-
min(axis=None, skipna=None, level=None, numeric_only=None, **kwargs)¶ Return the minimum of the values in the DataFrame.
- Parameters
- axis: {index (0), columns(1)}
Axis for the function to be applied on.
- skipna: bool, default True
Exclude NA/null values when computing the result.
- level: int or level name, default None
If the axis is a MultiIndex (hierarchical), count along a particular level, collapsing into a Series.
- numeric_only: bool, default None
Include only float, int, boolean columns. If None, will attempt to use everything, then use only numeric data.
- Returns
- Series
Notes
Parameters currently not supported are level, numeric_only.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.min() a 1 b 7 dtype: int64
-
mod(other, axis='columns', level=None, fill_value=None)¶ Get Modulo division of dataframe and other, element-wise (binary operator mod).
Equivalent to
dataframe % other, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, rmod.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df % 100 angles degrees circle 0 60 triangle 3 80 rectangle 4 60 >>> df.mod(100) angles degrees circle 0 60 triangle 3 80 rectangle 4 60
-
mode(axis=0, numeric_only=False, dropna=True)¶ Get the mode(s) of each element along the selected axis.
The mode of a set of values is the value that appears most often. It can be multiple values.
- Parameters
- axis{0 or ‘index’, 1 or ‘columns’}, default 0
The axis to iterate over while searching for the mode:
0 or ‘index’ : get mode of each column
1 or ‘columns’ : get mode of each row.
- numeric_onlybool, default False
If True, only apply to numeric columns.
- dropnabool, default True
Don’t consider counts of NA/NaN/NaT.
- Returns
- DataFrame
The modes of each column or row.
See also
cudf.core.series.Series.modeReturn the highest frequency value in a Series.
cudf.core.series.Series.value_countsReturn the counts of values in a Series.
Notes
axisparameter is currently not supported.Examples
>>> import cudf >>> df = cudf.DataFrame({ ... "species": ["bird", "mammal", "arthropod", "bird"], ... "legs": [2, 4, 8, 2], ... "wings": [2.0, None, 0.0, None] ... }) >>> df species legs wings 0 bird 2 2.0 1 mammal 4 <NA> 2 arthropod 8 0.0 3 bird 2 <NA>
By default, missing values are not considered, and the mode of wings are both 0 and 2. The second row of species and legs contains
NA, because they have only one mode, but the DataFrame has two rows.>>> df.mode() species legs wings 0 bird 2 0.0 1 <NA> <NA> 2.0
Setting
dropna=False,NAvalues are considered and they can be the mode (like for wings).>>> df.mode(dropna=False) species legs wings 0 bird 2 <NA>
Setting
numeric_only=True, only the mode of numeric columns is computed, and columns of other types are ignored.>>> df.mode(numeric_only=True) legs wings 0 2 0.0 1 <NA> 2.0
-
mul(other, axis='columns', level=None, fill_value=None)¶ Get Multiplication of dataframe and other, element-wise (binary operator mul).
Equivalent to
dataframe * other, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, rmul.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> other = cudf.DataFrame({'angles': [0, 3, 4]}, ... index=['circle', 'triangle', 'rectangle']) >>> df * other angles degrees circle 0 <NA> triangle 9 <NA> rectangle 16 <NA> >>> df.mul(other, fill_value=0) angles degrees circle 0 0 triangle 9 0 rectangle 16 0
-
nans_to_nulls()¶ Convert nans (if any) to nulls.
-
property
ndim¶ Dimension of the data. DataFrame ndim is always 2.
-
nlargest(n, columns, keep='first')¶ Get the rows of the DataFrame sorted by the n largest value of columns
Notes
- Difference from pandas:
Only a single column is supported in columns
-
notna()¶ Identify non-missing values.
Return a boolean same-sized object indicating if the values are not
<NA>. Non-missing values get mapped toTrue.<NA>values get mapped toFalsevalues.<NA>values include:Values where null mask is set.
NaNin float dtype.NaTin datetime64 and timedelta64 types.
Characters such as empty strings
''orinfincase of float are not considered<NA>values.- Returns
- DataFrame/Series/Index
Mask of bool values for each element in the object that indicates whether an element is not an NA value.
Examples
Show which entries in a DataFrame are NA.
>>> import cudf >>> import numpy as np >>> import pandas as pd >>> df = cudf.DataFrame({'age': [5, 6, np.NaN], ... 'born': [pd.NaT, pd.Timestamp('1939-05-27'), ... pd.Timestamp('1940-04-25')], ... 'name': ['Alfred', 'Batman', ''], ... 'toy': [None, 'Batmobile', 'Joker']}) >>> df age born name toy 0 5 <NA> Alfred <NA> 1 6 1939-05-27 00:00:00.000000 Batman Batmobile 2 <NA> 1940-04-25 00:00:00.000000 Joker >>> df.notnull() age born name toy 0 True False True False 1 True True True True 2 False True True True
Show which entries in a Series are NA.
>>> ser = cudf.Series([5, 6, np.NaN, np.inf, -np.inf]) >>> ser 0 5.0 1 6.0 2 <NA> 3 Inf 4 -Inf dtype: float64 >>> ser.notnull() 0 True 1 True 2 False 3 True 4 True dtype: bool
Show which entries in an Index are NA.
>>> idx = cudf.Index([1, 2, None, np.NaN, 0.32, np.inf]) >>> idx Float64Index([1.0, 2.0, <NA>, <NA>, 0.32, Inf], dtype='float64') >>> idx.notnull() GenericIndex([True, True, False, False, True, True], dtype='bool')
-
notnull()¶ Identify non-missing values.
Return a boolean same-sized object indicating if the values are not
<NA>. Non-missing values get mapped toTrue.<NA>values get mapped toFalsevalues.<NA>values include:Values where null mask is set.
NaNin float dtype.NaTin datetime64 and timedelta64 types.
Characters such as empty strings
''orinfincase of float are not considered<NA>values.- Returns
- DataFrame/Series/Index
Mask of bool values for each element in the object that indicates whether an element is not an NA value.
Examples
Show which entries in a DataFrame are NA.
>>> import cudf >>> import numpy as np >>> import pandas as pd >>> df = cudf.DataFrame({'age': [5, 6, np.NaN], ... 'born': [pd.NaT, pd.Timestamp('1939-05-27'), ... pd.Timestamp('1940-04-25')], ... 'name': ['Alfred', 'Batman', ''], ... 'toy': [None, 'Batmobile', 'Joker']}) >>> df age born name toy 0 5 <NA> Alfred <NA> 1 6 1939-05-27 00:00:00.000000 Batman Batmobile 2 <NA> 1940-04-25 00:00:00.000000 Joker >>> df.notnull() age born name toy 0 True False True False 1 True True True True 2 False True True True
Show which entries in a Series are NA.
>>> ser = cudf.Series([5, 6, np.NaN, np.inf, -np.inf]) >>> ser 0 5.0 1 6.0 2 <NA> 3 Inf 4 -Inf dtype: float64 >>> ser.notnull() 0 True 1 True 2 False 3 True 4 True dtype: bool
Show which entries in an Index are NA.
>>> idx = cudf.Index([1, 2, None, np.NaN, 0.32, np.inf]) >>> idx Float64Index([1.0, 2.0, <NA>, <NA>, 0.32, Inf], dtype='float64') >>> idx.notnull() GenericIndex([True, True, False, False, True, True], dtype='bool')
-
nsmallest(n, columns, keep='first')¶ Get the rows of the DataFrame sorted by the n smallest value of columns
Notes
- Difference from pandas:
Only a single column is supported in columns
-
one_hot_encoding(column, prefix, cats, prefix_sep='_', dtype='float64')¶ Expand a column with one-hot-encoding.
- Parameters
- columnstr
the source column with binary encoding for the data.
- prefixstr
the new column name prefix.
- catssequence of ints
the sequence of categories as integers.
- prefix_sepstr
the separator between the prefix and the category.
- dtype :
the dtype for the outputs; defaults to float64.
- Returns
- a new dataframe with new columns append for each category.
Examples
>>> import pandas as pd >>> import cudf >>> pet_owner = [1, 2, 3, 4, 5] >>> pet_type = ['fish', 'dog', 'fish', 'bird', 'fish'] >>> df = pd.DataFrame({'pet_owner': pet_owner, 'pet_type': pet_type}) >>> df.pet_type = df.pet_type.astype('category')
Create a column with numerically encoded category values
>>> df['pet_codes'] = df.pet_type.cat.codes >>> gdf = cudf.from_pandas(df)
Create the list of category codes to use in the encoding
>>> codes = gdf.pet_codes.unique() >>> gdf.one_hot_encoding('pet_codes', 'pet_dummy', codes).head() pet_owner pet_type pet_codes pet_dummy_0 pet_dummy_1 pet_dummy_2 0 1 fish 2 0.0 0.0 1.0 1 2 dog 1 0.0 1.0 0.0 2 3 fish 2 0.0 0.0 1.0 3 4 bird 0 1.0 0.0 0.0 4 5 fish 2 0.0 0.0 1.0
-
partition_by_hash(columns, nparts, keep_index=True)¶ Partition the dataframe by the hashed value of data in columns.
- Parameters
- columnssequence of str
The names of the columns to be hashed. Must have at least one name.
- npartsint
Number of output partitions
- keep_indexboolean
Whether to keep the index or drop it
- Returns
- partitioned: list of DataFrame
-
pipe(func, *args, **kwargs)¶ Apply
func(self, *args, **kwargs).- Parameters
- funcfunction
Function to apply to the Series/DataFrame/Index.
args, andkwargsare passed intofunc. Alternatively a(callable, data_keyword)tuple wheredata_keywordis a string indicating the keyword ofcallablethat expects the Series/DataFrame/Index.- argsiterable, optional
Positional arguments passed into
func.- kwargsmapping, optional
A dictionary of keyword arguments passed into
func.
- Returns
- objectthe return type of
func.
- objectthe return type of
Examples
Use
.pipewhen chaining together functions that expect Series, DataFrames or GroupBy objects. Instead of writing>>> func(g(h(df), arg1=a), arg2=b, arg3=c)
You can write
>>> (df.pipe(h) ... .pipe(g, arg1=a) ... .pipe(func, arg2=b, arg3=c) ... )
If you have a function that takes the data as (say) the second argument, pass a tuple indicating which keyword expects the data. For example, suppose
ftakes its data asarg2:>>> (df.pipe(h) ... .pipe(g, arg1=a) ... .pipe((func, 'arg2'), arg1=a, arg3=c) ... )
-
pivot(index, columns, values=None)¶ Return reshaped DataFrame organized by the given index and column values.
Reshape data (produce a “pivot” table) based on column values. Uses unique values from specified index / columns to form axes of the resulting DataFrame.
- Parameters
- indexcolumn name, optional
Column used to construct the index of the result.
- columnscolumn name, optional
Column used to construct the columns of the result.
- valuescolumn name or list of column names, optional
Column(s) whose values are rearranged to produce the result. If not specified, all remaining columns of the DataFrame are used.
- Returns
- DataFrame
Examples
>>> a = cudf.DataFrame() >>> a['a'] = [1, 1, 2, 2], >>> a['b'] = ['a', 'b', 'a', 'b'] >>> a['c'] = [1, 2, 3, 4] >>> a.pivot(index='a', columns='b') c b a b a 1 1 2 2 3 4
Pivot with missing values in result:
>>> a = cudf.DataFrame() >>> a['a'] = [1, 1, 2] >>> a['b'] = [1, 2, 3] >>> a['c'] = ['one', 'two', 'three'] >>> a.pivot(index='a', columns='b') c b 1 2 3 a 1 one two <NA> 2 <NA> <NA> three
-
pop(item)¶ Return a column and drop it from the DataFrame.
-
pow(other, axis='columns', level=None, fill_value=None)¶ Get Exponential power of dataframe and other, element-wise (binary operator pow).
Equivalent to
dataframe ** other, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, rpow.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [1, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df ** 2 angles degrees circle 0 129600 triangle 9 32400 rectangle 16 129600 >>> df.pow(2) angles degrees circle 0 129600 triangle 9 32400 rectangle 16 129600
-
prod(axis=None, skipna=None, dtype=None, level=None, numeric_only=None, min_count=0, **kwargs)¶ Return product of the values in the DataFrame.
- Parameters
- axis: {index (0), columns(1)}
Axis for the function to be applied on.
- skipna: bool, default True
Exclude NA/null values when computing the result.
- dtype: data type
Data type to cast the result to.
- min_count: int, default 0
The required number of valid values to perform the operation. If fewer than min_count non-NA values are present the result will be NA.
The default being 0. This means the sum of an all-NA or empty Series is 0, and the product of an all-NA or empty Series is 1.
- Returns
- scalar
Notes
Parameters currently not supported are level, numeric_only.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.prod() a 24 b 5040 dtype: int64
-
product(axis=None, skipna=None, dtype=None, level=None, numeric_only=None, min_count=0, **kwargs)¶ Return product of the values in the DataFrame.
- Parameters
- axis: {index (0), columns(1)}
Axis for the function to be applied on.
- skipna: bool, default True
Exclude NA/null values when computing the result.
- dtype: data type
Data type to cast the result to.
- min_count: int, default 0
The required number of valid values to perform the operation. If fewer than min_count non-NA values are present the result will be NA.
The default being 0. This means the sum of an all-NA or empty Series is 0, and the product of an all-NA or empty Series is 1.
- Returns
- Series
Notes
Parameters currently not supported are level`, numeric_only.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.product() a 24 b 5040 dtype: int64
-
quantile(q=0.5, axis=0, numeric_only=True, interpolation='linear', columns=None, exact=True)¶ Return values at the given quantile.
- Parameters
- qfloat or array-like
0 <= q <= 1, the quantile(s) to compute
- axisint
axis is a NON-FUNCTIONAL parameter
- numeric_onlybool, default True
If False, the quantile of datetime and timedelta data will be computed as well.
- interpolation{linear, lower, higher, midpoint, nearest}
This parameter specifies the interpolation method to use, when the desired quantile lies between two data points i and j. Default
linear.- columnslist of str
List of column names to include.
- exactboolean
Whether to use approximate or exact quantile algorithm.
- Returns
- Series or DataFrame
If q is an array or numeric_only is set to False, a DataFrame will be returned where index is q, the columns are the columns of self, and the values are the quantile.
If q is a float, a Series will be returned where the index is the columns of self and the values are the quantiles.
Notes
One notable difference from Pandas is when DataFrame is of non-numeric types and result is expected to be a Series in case of Pandas. cuDF will return a DataFrame as it doesn’t support mixed types under Series.
-
quantiles(q=0.5, interpolation='nearest')¶ Return values at the given quantile.
- Parameters
- qfloat or array-like
0 <= q <= 1, the quantile(s) to compute
- interpolation{lower, higher, nearest}
This parameter specifies the interpolation method to use, when the desired quantile lies between two data points i and j. Default ‘nearest’.
- Returns
- DataFrame
-
query(expr, local_dict=None)¶ Query with a boolean expression using Numba to compile a GPU kernel.
See pandas.DataFrame.query.
- Parameters
- exprstr
A boolean expression. Names in expression refer to columns. index can be used instead of index name, but this is not supported for MultiIndex.
Names starting with @ refer to Python variables.
An output value will be null if any of the input values are null regardless of expression.
- local_dictdict
Containing the local variable to be used in query.
- Returns
- filteredDataFrame
Examples
>>> import cudf >>> a = ('a', [1, 2, 2]) >>> b = ('b', [3, 4, 5]) >>> df = cudf.DataFrame([a, b]) >>> expr = "(a == 2 and b == 4) or (b == 3)" >>> df.query(expr) a b 0 1 3 1 2 4
DateTime conditionals:
>>> import numpy as np >>> import datetime >>> df = cudf.DataFrame() >>> data = np.array(['2018-10-07', '2018-10-08'], dtype='datetime64') >>> df['datetimes'] = data >>> search_date = datetime.datetime.strptime('2018-10-08', '%Y-%m-%d') >>> df.query('datetimes==@search_date') datetimes 1 2018-10-08T00:00:00.000
Using local_dict:
>>> import numpy as np >>> import datetime >>> df = cudf.DataFrame() >>> data = np.array(['2018-10-07', '2018-10-08'], dtype='datetime64') >>> df['datetimes'] = data >>> search_date2 = datetime.datetime.strptime('2018-10-08', '%Y-%m-%d') >>> df.query('datetimes==@search_date', ... local_dict={'search_date':search_date2}) datetimes 1 2018-10-08T00:00:00.000
-
radd(other, axis=1, level=None, fill_value=None)¶ Get Addition of dataframe and other, element-wise (binary operator radd).
Equivalent to
other + dataframe, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, add.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df + 1 angles degrees circle 1 361 triangle 4 181 rectangle 5 361 >>> df.radd(1) angles degrees circle 1 361 triangle 4 181 rectangle 5 361
-
rank(axis=0, method='average', numeric_only=None, na_option='keep', ascending=True, pct=False)¶ Compute numerical data ranks (1 through n) along axis. By default, equal values are assigned a rank that is the average of the ranks of those values.
- Parameters
- axis{0 or ‘index’, 1 or ‘columns’}, default 0
Index to direct ranking.
- method{‘average’, ‘min’, ‘max’, ‘first’, ‘dense’}, default ‘average’
How to rank the group of records that have the same value (i.e. ties): * average: average rank of the group * min: lowest rank in the group * max: highest rank in the group * first: ranks assigned in order they appear in the array * dense: like ‘min’, but rank always increases by 1 between groups.
- numeric_onlybool, optional
For DataFrame objects, rank only numeric columns if set to True.
- na_option{‘keep’, ‘top’, ‘bottom’}, default ‘keep’
How to rank NaN values: * keep: assign NaN rank to NaN values * top: assign smallest rank to NaN values if ascending * bottom: assign highest rank to NaN values if ascending.
- ascendingbool, default True
Whether or not the elements should be ranked in ascending order.
- pctbool, default False
Whether or not to display the returned rankings in percentile form.
- Returns
- same type as caller
Return a Series or DataFrame with data ranks as values.
-
rdiv(other, axis='columns', level=None, fill_value=None)¶ Get Floating division of dataframe and other, element-wise (binary operator rtruediv).
Equivalent to
other / dataframe, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, truediv.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df angles degrees circle 0 360 triangle 3 180 rectangle 4 360 >>> df.rtruediv(10) angles degrees circle inf 0.027778 triangle 3.333333 0.055556 rectangle 2.500000 0.027778 >>> df.rdiv(10) angles degrees circle inf 0.027778 triangle 3.333333 0.055556 rectangle 2.500000 0.027778 >>> 10 / df angles degrees circle inf 0.027778 triangle 3.333333 0.055556 rectangle 2.500000 0.027778
-
reindex(labels=None, axis=0, index=None, columns=None, copy=True)¶ Return a new DataFrame whose axes conform to a new index
DataFrame.reindexsupports two calling conventions:(index=index_labels, columns=column_names)(labels, axis={0 or 'index', 1 or 'columns'})
- Parameters
- labelsIndex, Series-convertible, optional, default None
- axis{0 or ‘index’, 1 or ‘columns’}, optional, default 0
- indexIndex, Series-convertible, optional, default None
Shorthand for
df.reindex(labels=index_labels, axis=0)- columnsarray-like, optional, default None
Shorthand for
df.reindex(labels=column_names, axis=1)- copyboolean, optional, default True
- Returns
- A DataFrame whose axes conform to the new index(es)
Examples
>>> import cudf >>> df = cudf.DataFrame() >>> df['key'] = [0, 1, 2, 3, 4] >>> df['val'] = [float(i + 10) for i in range(5)] >>> df_new = df.reindex(index=[0, 3, 4, 5], ... columns=['key', 'val', 'sum']) >>> df key val 0 0 10.0 1 1 11.0 2 2 12.0 3 3 13.0 4 4 14.0 >>> df_new key val sum 0 0 10.0 NaN 3 3 13.0 NaN 4 4 14.0 NaN 5 -1 NaN NaN
-
rename(mapper=None, index=None, columns=None, axis=0, copy=True, inplace=False, level=None, errors='ignore')¶ Alter column and index labels.
Function / dict values must be unique (1-to-1). Labels not contained in a dict / Series will be left as-is. Extra labels listed don’t throw an error.
DataFrame.renamesupports two calling conventions:(index=index_mapper, columns=columns_mapper, ...)(mapper, axis={0/'index' or 1/'column'}, ...)
We highly recommend using keyword arguments to clarify your intent.
- Parameters
- mapperdict-like or function, default None
optional dict-like or functions transformations to apply to the index/column values depending on selected
axis.- indexdict-like, default None
Optional dict-like transformations to apply to the index axis’ values. Does not support functions for axis 0 yet.
- columnsdict-like or function, default None
optional dict-like or functions transformations to apply to the columns axis’ values.
- axisint, default 0
Axis to rename with mapper. 0 or ‘index’ for index 1 or ‘columns’ for columns
- copyboolean, default True
Also copy underlying data
- inplaceboolean, default False
Return new DataFrame. If True, assign columns without copy
- levelint or level name, default None
In case of a MultiIndex, only rename labels in the specified level.
- errors{‘raise’, ‘ignore’, ‘warn’}, default ‘ignore’
Only ‘ignore’ supported Control raising of exceptions on invalid data for provided dtype.
raise: allow exceptions to be raisedignore: suppress exceptions. On error return original object.warn: prints last exceptions as warnings and return original object.
- Returns
- DataFrame
Notes
- Difference from pandas:
Not supporting: level
Rename will not overwite column names. If a list with duplicates is passed, column names will be postfixed with a number.
-
repeat(repeats, axis=None)¶ Repeats elements consecutively.
Returns a new object of caller type(DataFrame/Series/Index) where each element of the current object is repeated consecutively a given number of times.
- Parameters
- repeatsint, or array of ints
The number of repetitions for each element. This should be a non-negative integer. Repeating 0 times will return an empty object.
- Returns
- Series/DataFrame/Index
A newly created object of same type as caller with repeated elements.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3], 'b': [10, 20, 30]}) >>> df a b 0 1 10 1 2 20 2 3 30 >>> df.repeat(3) a b 0 1 10 0 1 10 0 1 10 1 2 20 1 2 20 1 2 20 2 3 30 2 3 30 2 3 30
Repeat on Series
>>> s = cudf.Series([0, 2]) >>> s 0 0 1 2 dtype: int64 >>> s.repeat([3, 4]) 0 0 0 0 0 0 1 2 1 2 1 2 1 2 dtype: int64 >>> s.repeat(2) 0 0 0 0 1 2 1 2 dtype: int64
Repeat on Index
>>> index = cudf.Index([10, 22, 33, 55]) >>> index Int64Index([10, 22, 33, 55], dtype='int64') >>> index.repeat(5) Int64Index([10, 10, 10, 10, 10, 22, 22, 22, 22, 22, 33, 33, 33, 33, 33, 55, 55, 55, 55, 55], dtype='int64')
-
replace(to_replace=None, value=None, inplace=False, limit=None, regex=False, method=None)¶ Replace values given in to_replace with replacement.
- Parameters
- to_replacenumeric, str, list-like or dict
Value(s) that will be replaced.
- numeric or str:
values equal to to_replace will be replaced with replacement
- list of numeric or str:
If replacement is also list-like, to_replace and replacement must be of same length.
- dict:
Dicts can be used to replace different values in different columns. For example, {‘a’: 1, ‘z’: 2} specifies that the value 1 in column a and the value 2 in column z should be replaced with replacement*.
Dicts can be used to specify different replacement values for different existing values. For example, {‘a’: ‘b’, ‘y’: ‘z’} replaces the value ‘a’ with ‘b’ and ‘y’ with ‘z’. To use a dict in this way the value parameter should be None.
- valuenumeric, str, list-like, or dict
Value(s) to replace to_replace with. If a dict is provided, then its keys must match the keys in to_replace, and corresponding values must be compatible (e.g., if they are lists, then they must match in length).
- inplacebool, default False
If True, in place.
- Returns
- resultDataFrame
DataFrame after replacement.
- Raises
- TypeError
If
to_replaceis not a scalar, array-like, dict, or NoneIf
to_replaceis a dict and value is not a list, dict, or Series
- ValueError
If a list is passed to
to_replaceandvaluebut they are not the same length.
Notes
Parameters that are currently not supported are: limit, regex, method
Examples
Scalar
to_replaceandvalue>>> import cudf >>> df = cudf.DataFrame({'A': [0, 1, 2, 3, 4], ... 'B': [5, 6, 7, 8, 9], ... 'C': ['a', 'b', 'c', 'd', 'e']}) >>> df A B C 0 0 5 a 1 1 6 b 2 2 7 c 3 3 8 d 4 4 9 e >>> df.replace(0, 5) A B C 0 5 5 a 1 1 6 b 2 2 7 c 3 3 8 d 4 4 9 e
List-like
to_replace>>> df.replace([0, 1, 2, 3], 4) A B C 0 4 5 a 1 4 6 b 2 4 7 c 3 4 8 d 4 4 9 e >>> df.replace([0, 1, 2, 3], [4, 3, 2, 1]) A B C 0 4 5 a 1 3 6 b 2 2 7 c 3 1 8 d 4 4 9 e
dict-like
to_replace>>> df.replace({0: 10, 1: 100}) A B C 0 10 5 a 1 100 6 b 2 2 7 c 3 3 8 d 4 4 9 e >>> df.replace({'A': 0, 'B': 5}, 100) A B C 0 100 100 a 1 1 6 b 2 2 7 c 3 3 8 d 4 4 9 e
-
reset_index(level=None, drop=False, inplace=False, col_level=0, col_fill='')¶ Reset the index.
Reset the index of the DataFrame, and use the default one instead.
- Parameters
- dropbool, default False
Do not try to insert index into dataframe columns. This resets the index to the default integer index.
- inplacebool, default False
Modify the DataFrame in place (do not create a new object).
- Returns
- DataFrame or None
DataFrame with the new index or None if
inplace=True.
Examples
>>> df = cudf.DataFrame([('bird', 389.0), ... ('bird', 24.0), ... ('mammal', 80.5), ... ('mammal', np.nan)], ... index=['falcon', 'parrot', 'lion', 'monkey'], ... columns=('class', 'max_speed')) >>> df class max_speed falcon bird 389.0 parrot bird 24.0 lion mammal 80.5 monkey mammal <NA> >>> df.reset_index() index class max_speed 0 falcon bird 389.0 1 parrot bird 24.0 2 lion mammal 80.5 3 monkey mammal <NA> >>> df.reset_index(drop=True) class max_speed 0 bird 389.0 1 bird 24.0 2 mammal 80.5 3 mammal <NA>
-
rfloordiv(other, axis='columns', level=None, fill_value=None)¶ Get Integer division of dataframe and other, element-wise (binary operator rfloordiv).
Equivalent to
other // dataframe, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, floordiv.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'col1': [10, 11, 23], ... 'col2': [101, 122, 321]}) >>> df col1 col2 0 10 101 1 11 122 2 23 321 >>> df.rfloordiv(df) col1 col2 0 1 1 1 1 1 2 1 1 >>> df.rfloordiv(200) col1 col2 0 20 1 1 18 1 2 8 0 >>> df.rfloordiv(100) col1 col2 0 10 0 1 9 0 2 4 0
-
rmod(other, axis='columns', level=None, fill_value=None)¶ Get Modulo division of dataframe and other, element-wise (binary operator rmod).
Equivalent to
other % dataframe, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, mod.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [1, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> 100 % df angles degrees circle 0 100 triangle 1 100 rectangle 0 100 >>> df.rmod(100) angles degrees circle 0 100 triangle 1 100 rectangle 0 100
-
rmul(other, axis='columns', level=None, fill_value=None)¶ Get Multiplication of dataframe and other, element-wise (binary operator rmul).
Equivalent to
other * dataframe, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, mul.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> other = cudf.DataFrame({'angles': [0, 3, 4]}, ... index=['circle', 'triangle', 'rectangle']) >>> other * df angles degrees circle 0 <NA> triangle 9 <NA> rectangle 16 <NA> >>> df.rmul(other, fill_value=0) angles degrees circle 0 0 triangle 9 0 rectangle 16 0
-
round(decimals=0)¶ Round a DataFrame to a variable number of decimal places.
- Parameters
- decimalsint, dict, Series
Number of decimal places to round each column to. If an int is given, round each column to the same number of places. Otherwise dict and Series round to variable numbers of places. Column names should be in the keys if decimals is a dict-like, or in the index if decimals is a Series. Any columns not included in decimals will be left as is. Elements of decimals which are not columns of the input will be ignored.
- Returns
- DataFrame
A DataFrame with the affected columns rounded to the specified number of decimal places.
Examples
>>> df = cudf.DataFrame( [(.21, .32), (.01, .67), (.66, .03), (.21, .18)], ... columns=['dogs', 'cats'] ... ) >>> df dogs cats 0 0.21 0.32 1 0.01 0.67 2 0.66 0.03 3 0.21 0.18
By providing an integer each column is rounded to the same number of decimal places
>>> df.round(1) dogs cats 0 0.2 0.3 1 0.0 0.7 2 0.7 0.0 3 0.2 0.2
With a dict, the number of places for specific columns can be specified with the column names as key and the number of decimal places as value
>>> df.round({'dogs': 1, 'cats': 0}) dogs cats 0 0.2 0.0 1 0.0 1.0 2 0.7 0.0 3 0.2 0.0
Using a Series, the number of places for specific columns can be specified with the column names as index and the number of decimal places as value
>>> decimals = cudf.Series([0, 1], index=['cats', 'dogs']) >>> df.round(decimals) dogs cats 0 0.2 0.0 1 0.0 1.0 2 0.7 0.0 3 0.2 0.0
-
rpow(other, axis='columns', level=None, fill_value=None)¶ Get Exponential power of dataframe and other, element-wise (binary operator pow).
Equivalent to
other ** dataframe, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, pow.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [1, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> 1 ** df angles degrees circle 1 1 triangle 1 1 rectangle 1 1 >>> df.rpow(1) angles degrees circle 1 1 triangle 1 1 rectangle 1 1
-
rsub(other, axis='columns', level=None, fill_value=None)¶ Get Subtraction of dataframe and other, element-wise (binary operator rsub).
Equivalent to
other - dataframe, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, sub.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df angles degrees circle 0 360 triangle 3 180 rectangle 4 360 >>> df.rsub(1) angles degrees circle 1 -359 triangle -2 -179 rectangle -3 -359 >>> df.rsub([1, 2]) angles degrees circle 1 -358 triangle -2 -178 rectangle -3 -358
-
rtruediv(other, axis='columns', level=None, fill_value=None)¶ Get Floating division of dataframe and other, element-wise (binary operator rtruediv).
Equivalent to
other / dataframe, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, truediv.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df angles degrees circle 0 360 triangle 3 180 rectangle 4 360 >>> df.rtruediv(10) angles degrees circle inf 0.027778 triangle 3.333333 0.055556 rectangle 2.500000 0.027778 >>> df.rdiv(10) angles degrees circle inf 0.027778 triangle 3.333333 0.055556 rectangle 2.500000 0.027778 >>> 10 / df angles degrees circle inf 0.027778 triangle 3.333333 0.055556 rectangle 2.500000 0.027778
-
sample(n=None, frac=None, replace=False, weights=None, random_state=None, axis=None, keep_index=True)¶ Return a random sample of items from an axis of object.
You can use random_state for reproducibility.
- Parameters
- nint, optional
Number of items from axis to return. Cannot be used with frac. Default = 1 if frac = None.
- fracfloat, optional
Fraction of axis items to return. Cannot be used with n.
- replacebool, default False
Allow or disallow sampling of the same row more than once. replace == True is not yet supported for axis = 1/”columns”
- weightsstr or ndarray-like, optional
Only supported for axis=1/”columns”
- random_stateint, numpy RandomState or None, default None
Seed for the random number generator (if int), or None. If None, a random seed will be chosen. if RandomState, seed will be extracted from current state.
- axis{0 or ‘index’, 1 or ‘columns’, None}, default None
Axis to sample. Accepts axis number or name. Default is stat axis for given data type (0 for Series and DataFrames). Series and Index doesn’t support axis=1.
- Returns
- Series or DataFrame or Index
A new object of same type as caller containing n items randomly sampled from the caller object.
Examples
>>> import cudf as cudf >>> df = cudf.DataFrame({"a":{1, 2, 3, 4, 5}}) >>> df.sample(3) a 1 2 3 4 0 1
>>> sr = cudf.Series([1, 2, 3, 4, 5]) >>> sr.sample(10, replace=True) 1 4 3 1 2 4 0 5 0 1 4 5 4 1 0 2 0 3 3 2 dtype: int64
>>> df = cudf.DataFrame( ... {"a":[1, 2], "b":[2, 3], "c":[3, 4], "d":[4, 5]}) >>> df.sample(2, axis=1) a c 0 1 3 1 2 4
-
scatter_by_map(map_index, map_size=None, keep_index=True, **kwargs)¶ Scatter to a list of dataframes.
Uses map_index to determine the destination of each row of the original DataFrame.
- Parameters
- map_indexSeries, str or list-like
Scatter assignment for each row
- map_sizeint
Length of output list. Must be >= uniques in map_index
- keep_indexbool
Conserve original index values for each row
- Returns
- A list of cudf.DataFrame objects.
-
searchsorted(values, side='left', ascending=True, na_position='last')¶ Find indices where elements should be inserted to maintain order
- Parameters
- valueFrame (Shape must be consistent with self)
Values to be hypothetically inserted into Self
- sidestr {‘left’, ‘right’} optional, default ‘left‘
If ‘left’, the index of the first suitable location found is given If ‘right’, return the last such index
- ascendingbool optional, default True
Sorted Frame is in ascending order (otherwise descending)
- na_positionstr {‘last’, ‘first’} optional, default ‘last‘
Position of null values in sorted order
- Returns
- 1-D cupy array of insertion points
Examples
>>> s = cudf.Series([1, 2, 3]) >>> s.searchsorted(4) 3 >>> s.searchsorted([0, 4]) array([0, 3], dtype=int32) >>> s.searchsorted([1, 3], side='left') array([0, 2], dtype=int32) >>> s.searchsorted([1, 3], side='right') array([1, 3], dtype=int32)
If the values are not monotonically sorted, wrong locations may be returned:
>>> s = cudf.Series([2, 1, 3]) >>> s.searchsorted(1) 0 # wrong result, correct would be 1
>>> df = cudf.DataFrame({'a': [1, 3, 5, 7], 'b': [10, 12, 14, 16]}) >>> df a b 0 1 10 1 3 12 2 5 14 3 7 16 >>> values_df = cudf.DataFrame({'a': [0, 2, 5, 6], ... 'b': [10, 11, 13, 15]}) >>> values_df a b 0 0 10 1 2 17 2 5 13 3 6 15 >>> df.searchsorted(values_df, ascending=False) array([4, 4, 4, 0], dtype=int32)
-
select_dtypes(include=None, exclude=None)¶ Return a subset of the DataFrame’s columns based on the column dtypes.
- Parameters
- includestr or list
which columns to include based on dtypes
- excludestr or list
which columns to exclude based on dtypes
- Returns
- DataFrame
The subset of the frame including the dtypes in
includeand excluding the dtypes inexclude.
- Raises
- ValueError
If both of
includeandexcludeare emptyIf
includeandexcludehave overlapping elements
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2] * 3, ... 'b': [True, False] * 3, ... 'c': [1.0, 2.0] * 3}) >>> df a b c 0 1 True 1.0 1 2 False 2.0 2 1 True 1.0 3 2 False 2.0 4 1 True 1.0 5 2 False 2.0 >>> df.select_dtypes(include='bool') b 0 True 1 False 2 True 3 False 4 True 5 False >>> df.select_dtypes(include=['float64']) c 0 1.0 1 2.0 2 1.0 3 2.0 4 1.0 5 2.0 >>> df.select_dtypes(exclude=['int']) b c 0 True 1.0 1 False 2.0 2 True 1.0 3 False 2.0 4 True 1.0 5 False 2.0
-
set_index(keys, drop=True, append=False, inplace=False, verify_integrity=False)¶ Return a new DataFrame with a new index
- Parameters
- keysIndex, Series-convertible, label-like, or list
Index : the new index. Series-convertible : values for the new index. Label-like : Label of column to be used as index. List : List of items from above.
- dropboolean, default True
Whether to drop corresponding column for str index argument
- appendboolean, default True
Whether to append columns to the existing index, resulting in a MultiIndex.
- inplaceboolean, default False
Modify the DataFrame in place (do not create a new object).
- verify_integrityboolean, default False
Check for duplicates in the new index.
Examples
>>> df = cudf.DataFrame({ ... "a": [1, 2, 3, 4, 5], ... "b": ["a", "b", "c", "d","e"], ... "c": [1.0, 2.0, 3.0, 4.0, 5.0] ... }) >>> df a b c 0 1 a 1.0 1 2 b 2.0 2 3 c 3.0 3 4 d 4.0 4 5 e 5.0
Set the index to become the ‘b’ column:
>>> df.set_index('b') a c b a 1 1.0 b 2 2.0 c 3 3.0 d 4 4.0 e 5 5.0
Create a MultiIndex using columns ‘a’ and ‘b’:
>>> df.set_index(["a", "b"]) c a b 1 a 1.0 2 b 2.0 3 c 3.0 4 d 4.0 5 e 5.0
Set new Index instance as index:
>>> df.set_index(cudf.RangeIndex(10, 15)) a b c 10 1 a 1.0 11 2 b 2.0 12 3 c 3.0 13 4 d 4.0 14 5 e 5.0
Setting append=True will combine current index with column a:
>>> df.set_index("a", append=True) b c a 0 1 a 1.0 1 2 b 2.0 2 3 c 3.0 3 4 d 4.0 4 5 e 5.0
set_index supports inplace parameter too:
>>> df.set_index("a", inplace=True) >>> df b c a 1 a 1.0 2 b 2.0 3 c 3.0 4 d 4.0 5 e 5.0
-
property
shape¶ Returns a tuple representing the dimensionality of the DataFrame.
-
shift(periods=1, freq=None, axis=0, fill_value=None)¶ Shift values by periods positions.
-
sin()¶ Get Trigonometric sine, element-wise.
- Returns
- DataFrame/Series/Index
Result of the trigonometric operation.
Examples
>>> import cudf >>> ser = cudf.Series([0.0, 0.32434, 0.5, 45, 90, 180, 360]) >>> ser 0 0.00000 1 0.32434 2 0.50000 3 45.00000 4 90.00000 5 180.00000 6 360.00000 dtype: float64 >>> ser.sin() 0 0.000000 1 0.318683 2 0.479426 3 0.850904 4 0.893997 5 -0.801153 6 0.958916 dtype: float64
sin operation on DataFrame:
>>> df = cudf.DataFrame({'first': [0.0, 5, 10, 15], ... 'second': [100.0, 360, 720, 300]}) >>> df first second 0 0.0 100.0 1 5.0 360.0 2 10.0 720.0 3 15.0 300.0 >>> df.sin() first second 0 0.000000 -0.506366 1 -0.958924 0.958916 2 -0.544021 -0.544072 3 0.650288 -0.999756
sin operation on Index:
>>> index = cudf.Index([-0.4, 100, -180, 90]) >>> index Float64Index([-0.4, 100.0, -180.0, 90.0], dtype='float64') >>> index.sin() Float64Index([-0.3894183423086505, -0.5063656411097588, 0.8011526357338306, 0.8939966636005579], dtype='float64')
-
property
size¶ Return the number of elements in the underlying data.
- Returns
- sizeSize of the DataFrame / Index / Series / MultiIndex
Examples
Size of an empty dataframe is 0.
>>> import cudf >>> df = cudf.DataFrame() >>> df Empty DataFrame Columns: [] Index: [] >>> df.size 0 >>> df = cudf.DataFrame(index=[1, 2, 3]) >>> df Empty DataFrame Columns: [] Index: [1, 2, 3] >>> df.size 0
DataFrame with values
>>> df = cudf.DataFrame({'a': [10, 11, 12], ... 'b': ['hello', 'rapids', 'ai']}) >>> df a b 0 10 hello 1 11 rapids 2 12 ai >>> df.size 6 >>> df.index RangeIndex(start=0, stop=3) >>> df.index.size 3
Size of an Index
>>> index = cudf.Index([]) >>> index Float64Index([], dtype='float64') >>> index.size 0 >>> index = cudf.Index([1, 2, 3, 10]) >>> index Int64Index([1, 2, 3, 10], dtype='int64') >>> index.size 4
Size of a MultiIndex
>>> midx = cudf.MultiIndex( ... levels=[["a", "b", "c", None], ["1", None, "5"]], ... codes=[[0, 0, 1, 2, 3], [0, 2, 1, 1, 0]], ... names=["x", "y"], ... ) >>> midx MultiIndex([( 'a', '1'), ( 'a', '5'), ( 'b', <NA>), ( 'c', <NA>), (<NA>, '1')], names=['x', 'y']) >>> midx.size 5
-
skew(axis=None, skipna=None, level=None, numeric_only=None, **kwargs)¶ Return unbiased Fisher-Pearson skew of a sample.
- Parameters
- skipna: bool, default True
Exclude NA/null values when computing the result.
- Returns
- Series
Notes
Parameters currently not supported are axis, level and numeric_only
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [3, 2, 3, 4], 'b': [7, 8, 10, 10]}) >>> df.skew() a 0.00000 b -0.37037 dtype: float64
-
sort_index(axis=0, level=None, ascending=True, inplace=False, kind=None, na_position='last', sort_remaining=True, ignore_index=False)¶ Sort object by labels (along an axis).
- Parameters
- axis{0 or ‘index’, 1 or ‘columns’}, default 0
The axis along which to sort. The value 0 identifies the rows, and 1 identifies the columns.
- levelint or level name or list of ints or list of level names
If not None, sort on values in specified index level(s). This is only useful in the case of MultiIndex.
- ascendingbool, default True
Sort ascending vs. descending.
- inplacebool, default False
If True, perform operation in-place.
- kindsorting method such as quick sort and others.
Not yet supported.
- na_position{‘first’, ‘last’}, default ‘last’
Puts NaNs at the beginning if first; last puts NaNs at the end.
- sort_remainingbool, default True
Not yet supported
- ignore_indexbool, default False
if True, index will be replaced with RangeIndex.
- Returns
- DataFrame or None
Examples
>>> df = cudf.DataFrame( ... {"b":[3, 2, 1], "a":[2, 1, 3]}, index=[1, 3, 2]) >>> df.sort_index(axis=0) b a 1 3 2 2 1 3 3 2 1 >>> df.sort_index(axis=1) a b 1 2 3 3 1 2 2 3 1
-
sort_values(by, axis=0, ascending=True, inplace=False, kind='quicksort', na_position='last', ignore_index=False)¶ Sort by the values row-wise.
- Parameters
- bystr or list of str
Name or list of names to sort by.
- ascendingbool or list of bool, default True
Sort ascending vs. descending. Specify list for multiple sort orders. If this is a list of bools, must match the length of the by.
- na_position{‘first’, ‘last’}, default ‘last’
‘first’ puts nulls at the beginning, ‘last’ puts nulls at the end
- ignore_indexbool, default False
If True, index will not be sorted.
- Returns
- sorted_objcuDF DataFrame
Notes
- Difference from pandas:
Support axis=’index’ only.
Not supporting: inplace, kind
Examples
>>> import cudf >>> a = ('a', [0, 1, 2]) >>> b = ('b', [-3, 2, 0]) >>> df = cudf.DataFrame([a, b]) >>> df.sort_values('b') a b 0 0 -3 2 2 0 1 1 2
-
sqrt()¶ Get the non-negative square-root of all elements, element-wise.
- Returns
- DataFrame/Series/Index
Result of the non-negative square-root of each element.
Examples
>>> import cudf >>> import cudf >>> ser = cudf.Series([10, 25, 81, 1.0, 100]) >>> ser 0 10.0 1 25.0 2 81.0 3 1.0 4 100.0 dtype: float64 >>> ser.sqrt() 0 3.162278 1 5.000000 2 9.000000 3 1.000000 4 10.000000 dtype: float64
sqrt operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-10.0, 100, 625], ... 'second': [1, 2, 0.4]}) >>> df first second 0 -10.0 1.0 1 100.0 2.0 2 625.0 0.4 >>> df.sqrt() first second 0 NaN 1.000000 1 10.0 1.414214 2 25.0 0.632456
sqrt operation on Index:
>>> index = cudf.Index([-10.0, 100, 625]) >>> index Float64Index([-10.0, 100.0, 625.0], dtype='float64') >>> index.sqrt() Float64Index([nan, 10.0, 25.0], dtype='float64')
-
stack(level=- 1, dropna=True)¶ Stack the prescribed level(s) from columns to index
Return a reshaped Series
- Parameters
- dropnabool, default True
Whether to drop rows in the resulting Series with missing values.
- Returns
- The stacked cudf.Series
Examples
>>> import cudf >>> df = cudf.DataFrame({'a':[0,1,3], 'b':[1,2,4]}) >>> df.stack() 0 a 0 b 1 1 a 1 b 2 2 a 3 b 4 dtype: int64
-
std(axis=None, skipna=None, level=None, ddof=1, numeric_only=None, **kwargs)¶ Return sample standard deviation of the DataFrame.
Normalized by N-1 by default. This can be changed using the ddof argument
- Parameters
- axis: {index (0), columns(1)}
Axis for the function to be applied on.
- skipna: bool, default True
Exclude NA/null values. If an entire row/column is NA, the result will be NA.
- ddof: int, default 1
Delta Degrees of Freedom. The divisor used in calculations is N - ddof, where N represents the number of elements.
- Returns
- Series
Notes
Parameters currently not supported are level and numeric_only
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.std() a 1.290994 b 1.290994 dtype: float64
-
sub(other, axis='columns', level=None, fill_value=None)¶ Get Subtraction of dataframe and other, element-wise (binary operator sub).
Equivalent to
dataframe - other, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, rsub.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df.sub(1) angles degrees circle -1 359 triangle 2 179 rectangle 3 359 >>> df.sub([1, 2]) angles degrees circle -1 358 triangle 2 178 rectangle 3 358
-
sum(axis=None, skipna=None, dtype=None, level=None, numeric_only=None, min_count=0, **kwargs)¶ Return sum of the values in the DataFrame.
- Parameters
- axis: {index (0), columns(1)}
Axis for the function to be applied on.
- skipna: bool, default True
Exclude NA/null values when computing the result.
- dtype: data type
Data type to cast the result to.
- min_count: int, default 0
The required number of valid values to perform the operation. If fewer than min_count non-NA values are present the result will be NA.
The default being 0. This means the sum of an all-NA or empty Series is 0, and the product of an all-NA or empty Series is 1.
- Returns
- Series
Notes
Parameters currently not supported are level, numeric_only.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.sum() a 10 b 34 dtype: int64
-
tail(n=5)¶ Returns the last n rows as a new DataFrame
Examples
>>> import cudf >>> df = cudf.DataFrame() >>> df['key'] = [0, 1, 2, 3, 4] >>> df['val'] = [float(i + 10) for i in range(5)] # insert column >>> df.tail(2) key val 3 3 13.0 4 4 14.0
-
take(positions, keep_index=True)¶ Return a new DataFrame containing the rows specified by positions
- Parameters
- positionsarray-like
Integer or boolean array-like specifying the rows of the output. If integer, each element represents the integer index of a row. If boolean, positions must be of the same length as self, and represents a boolean mask.
- Returns
- outDataFrame
New DataFrame
Examples
>>> a = cudf.DataFrame({'a': [1.0, 2.0, 3.0], ... 'b': cudf.Series(['a', 'b', 'c'])}) >>> a.take([0, 2, 2]) a b 0 1.0 a 2 3.0 c 2 3.0 c >>> a.take([True, False, True]) a b 0 1.0 a 2 3.0 c
-
tan()¶ Get Trigonometric tangent, element-wise.
- Returns
- DataFrame/Series/Index
Result of the trigonometric operation.
Examples
>>> import cudf >>> ser = cudf.Series([0.0, 0.32434, 0.5, 45, 90, 180, 360]) >>> ser 0 0.00000 1 0.32434 2 0.50000 3 45.00000 4 90.00000 5 180.00000 6 360.00000 dtype: float64 >>> ser.tan() 0 0.000000 1 0.336213 2 0.546302 3 1.619775 4 -1.995200 5 1.338690 6 -3.380140 dtype: float64
tan operation on DataFrame:
>>> df = cudf.DataFrame({'first': [0.0, 5, 10, 15], ... 'second': [100.0, 360, 720, 300]}) >>> df first second 0 0.0 100.0 1 5.0 360.0 2 10.0 720.0 3 15.0 300.0 >>> df.tan() first second 0 0.000000 -0.587214 1 -3.380515 -3.380140 2 0.648361 0.648446 3 -0.855993 45.244742
tan operation on Index:
>>> index = cudf.Index([-0.4, 100, -180, 90]) >>> index Float64Index([-0.4, 100.0, -180.0, 90.0], dtype='float64') >>> index.tan() Float64Index([-0.4227932187381618, -0.587213915156929, -1.3386902103511544, -1.995200412208242], dtype='float64')
-
tile(count)¶ Repeats the rows from self DataFrame count times to form a new DataFrame.
- Parameters
- selfinput Table containing columns to interleave.
- countNumber of times to tile “rows”. Must be non-negative.
- Returns
- The table containing the tiled “rows”.
Examples
>>> df = Dataframe([[8, 4, 7], [5, 2, 3]]) >>> count = 2 >>> df.tile(df, count) 0 1 2 0 8 4 7 1 5 2 3 0 8 4 7 1 5 2 3
-
to_arrow(preserve_index=True)¶ Convert to a PyArrow Table.
- Parameters
- preserve_indexbool, default True
whether index column and its meta data needs to be saved or not
- Returns
- PyArrow Table
Examples
>>> import cudf >>> df = cudf.DataFrame( ... {"a":[1, 2, 3], "b":[4, 5, 6]}, index=[1, 2, 3]) >>> df.to_arrow() pyarrow.Table a: int64 b: int64 index: int64 >>> df.to_arrow(preserve_index=False) pyarrow.Table a: int64 b: int64
-
to_csv(path_or_buf=None, sep=',', na_rep='', columns=None, header=True, index=True, line_terminator='\n', chunksize=None, encoding=None, compression=None, **kwargs)¶ Write a dataframe to csv file format.
- Parameters
- path_or_bufstr or file handle, default None
File path or object, if None is provided the result is returned as a string.
- sepchar, default ‘,’
Delimiter to be used.
- na_repstr, default ‘’
String to use for null entries
- columnslist of str, optional
Columns to write
- headerbool, default True
Write out the column names
- indexbool, default True
Write out the index as a column
- line_terminatorchar, default ‘n’
- chunksizeint or None, default None
Rows to write at a time
- encoding: str, default ‘utf-8’
A string representing the encoding to use in the output file Only ‘utf-8’ is currently supported
- compression: str, None
A string representing the compression scheme to use in the the output file Compression while writing csv is not supported currently
- Returns
- ——-
- None or str
If path_or_buf is None, returns the resulting csv format as a string. Otherwise returns None.
See also
Notes
Follows the standard of Pandas csv.QUOTE_NONNUMERIC for all output.
If to_csv leads to memory errors consider setting the chunksize argument.
Examples
Write a dataframe to csv.
>>> import cudf >>> filename = 'foo.csv' >>> df = cudf.DataFrame({'x': [0, 1, 2, 3], 'y': [1.0, 3.3, 2.2, 4.4], 'z': ['a', 'b', 'c', 'd']}) >>> df = df.set_index([3, 2, 1, 0]) >>> df.to_csv(filename)
-
to_dlpack()¶ Converts a cuDF object into a DLPack tensor.
DLPack is an open-source memory tensor structure: dmlc/dlpack.
This function takes a cuDF object and converts it to a PyCapsule object which contains a pointer to a DLPack tensor. This function deep copies the data into the DLPack tensor from the cuDF object.
- Parameters
- cudf_objDataFrame, Series, Index, or Column
- Returns
- pycapsule_objPyCapsule
Output DLPack tensor pointer which is encapsulated in a PyCapsule object.
-
to_feather(path, *args, **kwargs)¶ Write a DataFrame to the feather format.
- Parameters
- pathstr
File path
See also
-
to_hdf(path_or_buf, key, *args, **kwargs)¶ Write the contained data to an HDF5 file using HDFStore.
Hierarchical Data Format (HDF) is self-describing, allowing an application to interpret the structure and contents of a file with no outside information. One HDF file can hold a mix of related objects which can be accessed as a group or as individual objects.
In order to add another DataFrame or Series to an existing HDF file please use append mode and a different a key.
For more information see the user guide.
- Parameters
- path_or_bufstr or pandas.HDFStore
File path or HDFStore object.
- keystr
Identifier for the group in the store.
- mode{‘a’, ‘w’, ‘r+’}, default ‘a’
Mode to open file:
‘w’: write, a new file is created (an existing file with the same name would be deleted).
‘a’: append, an existing file is opened for reading and writing, and if the file does not exist it is created.
‘r+’: similar to ‘a’, but the file must already exist.
- format{‘fixed’, ‘table’}, default ‘fixed’
Possible values:
‘fixed’: Fixed format. Fast writing/reading. Not-appendable, nor searchable.
‘table’: Table format. Write as a PyTables Table structure which may perform worse but allow more flexible operations like searching / selecting subsets of the data.
- appendbool, default False
For Table formats, append the input data to the existing.
- data_columnslist of columns or True, optional
List of columns to create as indexed data columns for on-disk queries, or True to use all columns. By default only the axes of the object are indexed. See Query via Data Columns. Applicable only to format=’table’.
- complevel{0-9}, optional
Specifies a compression level for data. A value of 0 disables compression.
- complib{‘zlib’, ‘lzo’, ‘bzip2’, ‘blosc’}, default ‘zlib’
Specifies the compression library to be used. As of v0.20.2 these additional compressors for Blosc are supported (default if no compressor specified: ‘blosc:blosclz’): {‘blosc:blosclz’, ‘blosc:lz4’, ‘blosc:lz4hc’, ‘blosc:snappy’, ‘blosc:zlib’, ‘blosc:zstd’}. Specifying a compression library which is not available issues a ValueError.
- fletcher32bool, default False
If applying compression use the fletcher32 checksum.
- dropnabool, default False
If true, ALL nan rows will not be written to store.
- errorsstr, default ‘strict’
Specifies how encoding and decoding errors are to be handled. See the errors argument for
open()for a full list of options.
See also
cudf.io.hdf.read_hdfRead from HDF file.
cudf.io.parquet.to_parquetWrite a DataFrame to the binary parquet format.
cudf.io.feather.to_featherWrite out feather-format for DataFrames.
-
to_json(path_or_buf=None, *args, **kwargs)¶ Convert the cuDF object to a JSON string. Note nulls and NaNs will be converted to null and datetime objects will be converted to UNIX timestamps.
- Parameters
- path_or_bufstring or file handle, optional
File path or object. If not specified, the result is returned as a string.
- orientstring
Indication of expected JSON string format.
- Series
default is ‘index’
allowed values are: {‘split’,’records’,’index’,’table’}
- DataFrame
default is ‘columns’
allowed values are: {‘split’,’records’,’index’,’columns’,’values’,’table’}
- The format of the JSON string
‘split’ : dict like {‘index’ -> [index], ‘columns’ -> [columns], ‘data’ -> [values]}
‘records’ : list like [{column -> value}, … , {column -> value}]
‘index’ : dict like {index -> {column -> value}}
‘columns’ : dict like {column -> {index -> value}}
‘values’ : just the values array
‘table’ : dict like {‘schema’: {schema}, ‘data’: {data}} describing the data, and the data component is like
orient='records'.
- date_format{None, ‘epoch’, ‘iso’}
Type of date conversion. ‘epoch’ = epoch milliseconds, ‘iso’ = ISO8601. The default depends on the orient. For
orient='table', the default is ‘iso’. For all other orients, the default is ‘epoch’.- double_precisionint, default 10
The number of decimal places to use when encoding floating point values.
- force_asciibool, default True
Force encoded string to be ASCII.
- date_unitstring, default ‘ms’ (milliseconds)
The time unit to encode to, governs timestamp and ISO8601 precision. One of ‘s’, ‘ms’, ‘us’, ‘ns’ for second, millisecond, microsecond, and nanosecond respectively.
- default_handlercallable, default None
Handler to call if object cannot otherwise be converted to a suitable format for JSON. Should receive a single argument which is the object to convert and return a serializable object.
- linesbool, default False
If ‘orient’ is ‘records’ write out line delimited json format. Will throw ValueError if incorrect ‘orient’ since others are not list like.
- compression{‘infer’, ‘gzip’, ‘bz2’, ‘zip’, ‘xz’, None}
A string representing the compression to use in the output file, only used when the first argument is a filename. By default, the compression is inferred from the filename.
- indexbool, default True
Whether to include the index values in the JSON string. Not including the index (
index=False) is only supported when orient is ‘split’ or ‘table’.
See also
-
to_orc(fname, compression=None, *args, **kwargs)¶ Write a DataFrame to the ORC format.
- Parameters
- fnamestr
File path or object where the ORC dataset will be stored.
- compression{{ ‘snappy’, None }}, default None
Name of the compression to use. Use None for no compression.
- enable_statistics: boolean, default True
Enable writing column statistics.
See also
-
to_pandas(nullable=False, **kwargs)¶ Convert to a Pandas DataFrame.
- Parameters
- nullableBoolean, Default False
If
nullableisTrue, the resulting columns in the dataframe will be having a corresponding nullable Pandas dtype. IfnullableisFalse, the resulting columns will either convert null values tonp.nanorNonedepending on the dtype.
- Returns
- outPandas DataFrame
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [0, 1, 2], 'b': [-3, 2, 0]}) >>> pdf = df.to_pandas() >>> pdf a b 0 0 -3 1 1 2 2 2 0 >>> type(pdf) <class 'pandas.core.frame.DataFrame'>
nullableparameter can be used to control whether dtype can be Pandas Nullable or not:>>> df = cudf.DataFrame({'a': [0, None, 2], 'b': [True, False, None]}) >>> df a b 0 0 True 1 <NA> False 2 2 <NA> >>> pdf = df.to_pandas(nullable=True) >>> pdf a b 0 0 True 1 <NA> False 2 2 <NA> >>> pdf.dtypes a Int64 b boolean dtype: object >>> pdf = df.to_pandas(nullable=False) >>> pdf a b 0 0.0 True 1 NaN False 2 2.0 None >>> pdf.dtypes a float64 b object dtype: object
-
to_parquet(path, *args, **kwargs)¶ Write a DataFrame to the parquet format.
- Parameters
- pathstr
File path or Root Directory path. Will be used as Root Directory path while writing a partitioned dataset.
- compression{‘snappy’, None}, default ‘snappy’
Name of the compression to use. Use
Nonefor no compression.- indexbool, default None
If
True, include the dataframe’s index(es) in the file output. IfFalse, they will not be written to the file. IfNone, the engine’s default behavior will be used.- partition_colslist, optional, default None
Column names by which to partition the dataset Columns are partitioned in the order they are given
- partition_file_namestr, optional, default None
File name to use for partitioned datasets. Different partitions will be written to different directories, but all files will have this name. If nothing is specified, a random uuid4 hex string will be used for each file.
- int96_timestampsbool, default False
If
True, write timestamps in int96 format. This will convert timestamps from timestamp[ns], timestamp[ms], timestamp[s], and timestamp[us] to the int96 format, which is the number of Julian days and the number of nanoseconds since midnight. IfFalse, timestamps will not be altered.
-
to_records(index=True)¶ Convert to a numpy recarray
- Parameters
- indexbool
Whether to include the index in the output.
- Returns
- numpy recarray
-
to_string()¶ Convert to string
cuDF uses Pandas internals for efficient string formatting. Set formatting options using pandas string formatting options and cuDF objects will print identically to Pandas objects.
cuDF supports null/None as a value in any column type, which is transparently supported during this output process.
Examples
>>> import cudf >>> df = cudf.DataFrame() >>> df['key'] = [0, 1, 2] >>> df['val'] = [float(i + 10) for i in range(3)] >>> df.to_string() ' key val\n0 0 10.0\n1 1 11.0\n2 2 12.0'
-
transpose()¶ Transpose index and columns.
- Returns
- a new (ncol x nrow) dataframe. self is (nrow x ncol)
Notes
Difference from pandas: Not supporting copy because default and only behavior is copy=True
-
truediv(other, axis='columns', level=None, fill_value=None)¶ Get Floating division of dataframe and other, element-wise (binary operator truediv).
Equivalent to
dataframe / other, but with support to substitute a fill_value for missing data in one of the inputs. With reverse version, rtruediv.Among flexible wrappers (add, sub, mul, div, mod, pow) to arithmetic operators: +, -, *, /, //, %, **.
- Parameters
- otherscalar, sequence, Series, or DataFrame
Any single or multiple element data structure, or list-like object.
- fill_valuefloat or None, default None
Fill existing missing (NaN) values, and any new element needed for successful DataFrame alignment, with this value before computation. If data in both corresponding DataFrame locations is missing the result will be missing.
- Returns
- DataFrame
Result of the arithmetic operation.
Examples
>>> import cudf >>> df = cudf.DataFrame({'angles': [0, 3, 4], ... 'degrees': [360, 180, 360]}, ... index=['circle', 'triangle', 'rectangle']) >>> df.truediv(10) angles degrees circle 0.0 36.0 triangle 0.3 18.0 rectangle 0.4 36.0 >>> df.div(10) angles degrees circle 0.0 36.0 triangle 0.3 18.0 rectangle 0.4 36.0 >>> df / 10 angles degrees circle 0.0 36.0 triangle 0.3 18.0 rectangle 0.4 36.0
-
unstack(level=- 1, fill_value=None)¶ Pivot one or more levels of the (necessarily hierarchical) index labels.
Pivots the specified levels of the index labels of df to the innermost levels of the columns labels of the result.
If the index of
dfhas multiple levels, returns aDataframewith specified level of the index pivoted to the column levels.If the index of
dfhas single level, returns aSerieswith all column levels pivoted to the index levels.
- Parameters
- dfDataFrame
- levellevel name or index, list-like
Integer, name or list of such, specifying one or more levels of the index to pivot
- fill_value
Non-functional argument provided for compatibility with Pandas.
- Returns
- Series or DataFrame
Examples
>>> df['a'] = [1, 1, 1, 2, 2] >>> df['b'] = [1, 2, 3, 1, 2] >>> df['c'] = [5, 6, 7, 8, 9] >>> df['d'] = ['a', 'b', 'a', 'd', 'e'] >>> df = df.set_index(['a', 'b', 'd']) >>> df c a b d 1 1 a 5 2 b 6 3 a 7 2 1 d 8 2 e 9
Unstacking level ‘a’:
>>> df.unstack('a') c a 1 2 b d 1 a 5 <NA> d <NA> 8 2 b 6 <NA> e <NA> 9 3 a 7 <NA>
Unstacking level ‘d’ :
>>> df.unstack('d') c d a b d e a b 1 1 5 <NA> <NA> <NA> 2 <NA> 6 <NA> <NA> 3 7 <NA> <NA> <NA> 2 1 <NA> <NA> 8 <NA> 2 <NA> <NA> <NA> 9
Unstacking multiple levels:
>>> df.unstack(['b', 'd']) c b 1 2 3 d a d b e a a 1 5 <NA> 6 <NA> 7 2 <NA> 8 <NA> 9 <NA>
Unstacking single level index dataframe:
>>> df = cudf.DataFrame({('c', 1): [1, 2, 3], ('c', 2):[9, 8, 7]}) >>> df.unstack() c 1 0 1 1 2 2 3 2 0 9 1 8 2 7 dtype: int64
-
update(other, join='left', overwrite=True, filter_func=None, errors='ignore')¶ Modify a DataFrame in place using non-NA values from another DataFrame.
Aligns on indices. There is no return value.
- Parameters
- otherDataFrame, or object coercible into a DataFrame
Should have at least one matching index/column label with the original DataFrame. If a Series is passed, its name attribute must be set, and that will be used as the column name to align with the original DataFrame.
- join{‘left’}, default ‘left’
Only left join is implemented, keeping the index and columns of the original object.
- overwrite{True, False}, default True
How to handle non-NA values for overlapping keys: True: overwrite original DataFrame’s values with values from other. False: only update values that are NA in the original DataFrame.
- filter_funcNone
filter_func is not supported yet Return True for values that should be updated.S
- errors{‘raise’, ‘ignore’}, default ‘ignore’
If ‘raise’, will raise a ValueError if the DataFrame and other both contain non-NA data in the same place.
- Returns
- Nonemethod directly changes calling object
- Raises
- ValueError
When
errors= ‘raise’ and there’s overlapping non-NA data.When
errorsis not either ‘ignore’ or ‘raise’
- NotImplementedError
If
join!= ‘left’
-
property
values¶ Return a CuPy representation of the DataFrame.
Only the values in the DataFrame will be returned, the axes labels will be removed.
- Returns
- out: cupy.ndarray
The values of the DataFrame.
-
var(axis=None, skipna=None, level=None, ddof=1, numeric_only=None, **kwargs)¶ Return unbiased variance of the DataFrame.
Normalized by N-1 by default. This can be changed using the ddof argument
- Parameters
- axis: {index (0), columns(1)}
Axis for the function to be applied on.
- skipna: bool, default True
Exclude NA/null values. If an entire row/column is NA, the result will be NA.
- ddof: int, default 1
Delta Degrees of Freedom. The divisor used in calculations is N - ddof, where N represents the number of elements.
- Returns
- scalar
Notes
Parameters currently not supported are level and numeric_only
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3, 4], 'b': [7, 8, 9, 10]}) >>> df.var() a 1.666667 b 1.666667 dtype: float64
-
where(cond, other=None, inplace=False)¶ Replace values where the condition is False.
- Parameters
- condbool Series/DataFrame, array-like
Where cond is True, keep the original value. Where False, replace with corresponding value from other. Callables are not supported.
- other: scalar, list of scalars, Series/DataFrame
Entries where cond is False are replaced with corresponding value from other. Callables are not supported. Default is None.
DataFrame expects only Scalar or array like with scalars or dataframe with same dimension as self.
Series expects only scalar or series like with same length
- inplacebool, default False
Whether to perform the operation in place on the data.
- Returns
- Same type as caller
Examples
>>> import cudf >>> df = cudf.DataFrame({"A":[1, 4, 5], "B":[3, 5, 8]}) >>> df.where(df % 2 == 0, [-1, -1]) A B 0 -1 -1 1 4 -1 2 -1 8
>>> ser = cudf.Series([4, 3, 2, 1, 0]) >>> ser.where(ser > 2, 10) 0 4 1 3 2 10 3 10 4 10 dtype: int64 >>> ser.where(ser > 2) 0 4 1 3 2 <NA> 3 <NA> 4 <NA> dtype: int64
Series¶
-
class
cudf.core.series.Series(data=None, index=None, dtype=None, name=None, nan_as_null=True)¶ One-dimensional GPU array (including time series).
Labels need not be unique but must be a hashable type. The object supports both integer- and label-based indexing and provides a host of methods for performing operations involving the index. Statistical methods from ndarray have been overridden to automatically exclude missing data (currently represented as null/NaN).
Operations between Series (+, -, /, *, **) align values based on their associated index values-– they need not be the same length. The result index will be the sorted union of the two indexes.
Seriesobjects are used as columns ofDataFrame.- Parameters
- dataarray-like, Iterable, dict, or scalar value
Contains data stored in Series.
- indexarray-like or Index (1d)
Values must be hashable and have the same length as data. Non-unique index values are allowed. Will default to RangeIndex (0, 1, 2, …, n) if not provided. If both a dict and index sequence are used, the index will override the keys found in the dict.
- dtypestr, numpy.dtype, or ExtensionDtype, optional
Data type for the output Series. If not specified, this will be inferred from data.
- namestr, optional
The name to give to the Series.
- nan_as_nullbool, Default True
If
None/True, convertsnp.nanvalues tonullvalues. IfFalse, leavesnp.nanvalues as is.
- Attributes
catAccessor object for categorical properties of the Series values.
dataThe gpu buffer for the data
dtAccessor object for datetimelike properties of the Series values.
dtypedtype of the Series
emptyIndicator whether DataFrame or Series is empty.
has_nullsIndicator whether Series contains null values.
ilocSelect values by position.
indexThe index object
is_monotonicReturn boolean if values in the object are monotonic_increasing.
is_monotonic_decreasingReturn boolean if values in the object are monotonic_decreasing.
is_monotonic_increasingReturn boolean if values in the object are monotonic_increasing.
is_uniqueReturn boolean if values in the object are unique.
- list
locSelect values by label.
nameReturns name of the Series.
ndimDimension of the data.
null_countNumber of null values
nullableA boolean indicating whether a null-mask is needed
nullmaskThe gpu buffer for the null-mask
shapeReturns a tuple representing the dimensionality of the Series.
sizeReturn the number of elements in the underlying data.
strVectorized string functions for Series and Index.
valid_countNumber of non-null values
valuesReturn a CuPy representation of the Series.
values_hostReturn a numpy representation of the Series.
Methods
abs()Absolute value of each element of the series.
acos()Get Trigonometric inverse cosine, element-wise.
add(other[, fill_value, axis])Addition of series and other, element-wise (binary operator add).
all([axis, bool_only, skipna, level])Return whether all elements are True in Series.
any([axis, bool_only, skipna, level])Return whether any elements is True in Series.
append(to_append[, ignore_index, …])Append values from another
Seriesor array-like object.applymap(udf[, out_dtype])Apply an elementwise function to transform the values in the Column.
argsort([ascending, na_position])Returns a Series of int64 index that will sort the series.
as_index()Returns a new Series with a RangeIndex.
as_mask()Convert booleans to bitmask
asin()Get Trigonometric inverse sine, element-wise.
astype(dtype[, copy, errors])Cast the Series to the given dtype
atan()Get Trigonometric inverse tangent, element-wise.
ceil()Rounds each value upward to the smallest integral value not less than the original.
clip([lower, upper, inplace, axis])Trim values at input threshold(s).
copy([deep])Make a copy of this object’s indices and data.
corr(other[, method, min_periods])Calculates the sample correlation between two Series, excluding missing values.
cos()Get Trigonometric cosine, element-wise.
count([level])Return number of non-NA/null observations in the Series
cov(other[, min_periods])Compute covariance with Series, excluding missing values.
cummax([axis, skipna])Return cumulative maximum of the Series.
cummin([axis, skipna])Return cumulative minimum of the Series.
cumprod([axis, skipna])Return cumulative product of the Series.
cumsum([axis, skipna])Return cumulative sum of the Series.
describe([percentiles, include, exclude, …])Generate descriptive statistics.
diff([periods])Calculate the difference between values at positions i and i - N in an array and store the output in a new array.
digitize(bins[, right])Return the indices of the bins to which each value in series belongs.
drop_duplicates([keep, inplace, ignore_index])Return Series with duplicate values removed
dropna([axis, inplace, how])Return a Series with null values removed.
eq(other[, fill_value, axis])Equal to of series and other, element-wise (binary operator eq).
equals(other, **kwargs)Test whether two objects contain the same elements.
exp()Get the exponential of all elements, element-wise.
factorize([na_sentinel])Encode the input values as integer labels
fillna([value, method, axis, inplace, limit])Fill null values with
valueor specifiedmethod.floor()Rounds each value downward to the largest integral value not greater than the original.
floordiv(other[, fill_value, axis])Integer division of series and other, element-wise (binary operator floordiv).
from_arrow(array)Convert from PyArrow Array/ChunkedArray to Series.
from_categorical(categorical[, codes])Creates from a pandas.Categorical
from_masked_array(data, mask[, null_count])Create a Series with null-mask.
from_pandas(s[, nan_as_null])Convert from a Pandas Series.
ge(other[, fill_value, axis])Greater than or equal to of series and other, element-wise (binary operator ge).
groupby([by, axis, level, as_index, sort, …])Group Series using a mapper or by a Series of columns.
gt(other[, fill_value, axis])Greater than of series and other, element-wise (binary operator gt).
hash_encode(stop[, use_name])Encode column values as ints in [0, stop) using hash function.
Compute the hash of values in this column.
head([n])Return the first n rows.
Interleave Series columns of a table into a single column.
isin(values)Check whether values are contained in Series.
isna()Identify missing values.
isnull()Identify missing values.
keys()Return alias for index.
kurt([axis, skipna, level, numeric_only])Return Fisher’s unbiased kurtosis of a sample.
kurtosis([axis, skipna, level, numeric_only])Return Fisher’s unbiased kurtosis of a sample.
label_encoding(cats[, dtype, na_sentinel])Perform label encoding
le(other[, fill_value, axis])Less than or equal to of series and other, element-wise (binary operator le).
log()Get the natural logarithm of all elements, element-wise.
lt(other[, fill_value, axis])Less than of series and other, element-wise (binary operator lt).
map(arg[, na_action])Map values of Series according to input correspondence.
mask(cond[, other, inplace])Replace values where the condition is True.
max([axis, skipna, dtype, level, numeric_only])Return the maximum of the values in the Series.
mean([axis, skipna, level, numeric_only])Return the mean of the values in the series.
median([axis, skipna, level, numeric_only])Return the median of the values for the requested axis.
memory_usage([index, deep])Return the memory usage of the Series.
min([axis, skipna, dtype, level, numeric_only])Return the minimum of the values in the Series.
mod(other[, fill_value, axis])Modulo of series and other, element-wise (binary operator mod).
mode([dropna])Return the mode(s) of the dataset.
mul(other[, fill_value, axis])Multiplication of series and other, element-wise (binary operator mul).
Convert nans (if any) to nulls
ne(other[, fill_value, axis])Not equal to of series and other, element-wise (binary operator ne).
nlargest([n, keep])Returns a new Series of the n largest element.
notna()Identify non-missing values.
notnull()Identify non-missing values.
nsmallest([n, keep])Returns a new Series of the n smallest element.
nunique([method, dropna])Returns the number of unique values of the Series: approximate version, and exact version to be moved to libgdf
one_hot_encoding(cats[, dtype])Perform one-hot-encoding
pipe(func, *args, **kwargs)Apply
func(self, *args, **kwargs).pow(other[, fill_value, axis])Exponential power of series and other, element-wise (binary operator pow).
prod([axis, skipna, dtype, level, …])Return product of the values in the series
product([axis, skipna, dtype, level, …])Return product of the values in the Series.
quantile([q, interpolation, exact, quant_index])Return values at the given quantile.
radd(other[, fill_value, axis])Addition of series and other, element-wise (binary operator radd).
rank([axis, method, numeric_only, …])Compute numerical data ranks (1 through n) along axis.
reindex([index, copy])Return a Series that conforms to a new index
rename([index, copy])Alter Series name
repeat(repeats[, axis])Repeats elements consecutively.
replace([to_replace, value, inplace, limit, …])Replace values given in
to_replacewithvalue.reset_index([drop, inplace])Reset index to RangeIndex
reverse()Reverse the Series
rfloordiv(other[, fill_value, axis])Integer division of series and other, element-wise (binary operator rfloordiv).
rmod(other[, fill_value, axis])Modulo of series and other, element-wise (binary operator rmod).
rmul(other[, fill_value, axis])Multiplication of series and other, element-wise (binary operator rmul).
rolling(window[, min_periods, center, axis, …])Rolling window calculations.
round([decimals])Round each value in a Series to the given number of decimals.
rpow(other[, fill_value, axis])Exponential power of series and other, element-wise (binary operator rpow).
rsub(other[, fill_value, axis])Subtraction of series and other, element-wise (binary operator rsub).
rtruediv(other[, fill_value, axis])Floating division of series and other, element-wise (binary operator rtruediv).
sample([n, frac, replace, weights, …])Return a random sample of items from an axis of object.
scale()Scale values to [0, 1] in float64
scatter_by_map(map_index[, map_size, keep_index])Scatter to a list of dataframes.
searchsorted(values[, side, ascending, …])Find indices where elements should be inserted to maintain order
set_index(index)Returns a new Series with a different index.
set_mask(mask[, null_count])Create new Series by setting a mask array.
shift([periods, freq, axis, fill_value])Shift values by periods positions.
sin()Get Trigonometric sine, element-wise.
skew([axis, skipna, level, numeric_only])Return unbiased Fisher-Pearson skew of a sample.
sort_index([ascending])Sort by the index.
sort_values([axis, ascending, inplace, …])Sort by the values.
sqrt()Get the non-negative square-root of all elements, element-wise.
std([axis, skipna, level, ddof, numeric_only])Return sample standard deviation of the Series.
sub(other[, fill_value, axis])Subtraction of series and other, element-wise (binary operator sub).
sum([axis, skipna, dtype, level, …])Return sum of the values in the Series.
tail([n])Returns the last n rows as a new Series
take(indices[, keep_index])Return Series by taking values from the corresponding indices.
tan()Get Trigonometric tangent, element-wise.
tile(count)Repeats the rows from self DataFrame count times to form a new DataFrame.
to_array([fillna])Get a dense numpy array for the data.
to_arrow()Convert Series to a PyArrow Array.
Converts a cuDF object into a DLPack tensor.
to_frame([name])Convert Series into a DataFrame
to_gpu_array([fillna])Get a dense numba device array for the data.
to_hdf(path_or_buf, key, *args, **kwargs)Write the contained data to an HDF5 file using HDFStore.
to_json([path_or_buf])Convert the cuDF object to a JSON string.
to_pandas([index, nullable])Convert to a Pandas Series.
Convert to string
truediv(other[, fill_value, axis])Floating division of series and other, element-wise (binary operator truediv).
unique()Returns unique values of this Series.
value_counts([normalize, sort, ascending, …])Return a Series containing counts of unique values.
values_to_string([nrows])Returns a list of string for each element.
var([axis, skipna, level, ddof, numeric_only])Return unbiased variance of the Series.
where(cond[, other, inplace])Replace values where the condition is False.
-
abs()¶ Absolute value of each element of the series.
Returns a new Series.
-
acos()¶ Get Trigonometric inverse cosine, element-wise.
The inverse of cos so that, if y = x.cos(), then x = y.acos()
- Returns
- DataFrame/Series/Index
Result of the trigonometric operation.
Examples
>>> import cudf >>> ser = cudf.Series([-1, 0, 1, 0.32434, 0.5]) >>> ser.acos() 0 3.141593 1 1.570796 2 0.000000 3 1.240482 4 1.047198 dtype: float64
acos operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-1, 0, 0.5], ... 'second': [0.234, 0.3, 0.1]}) >>> df first second 0 -1.0 0.234 1 0.0 0.300 2 0.5 0.100 >>> df.acos() first second 0 3.141593 1.334606 1 1.570796 1.266104 2 1.047198 1.470629
acos operation on Index:
>>> index = cudf.Index([-1, 0.4, 1, 0, 0.3]) >>> index Float64Index([-1.0, 0.4, 1.0, 0.0, 0.3], dtype='float64') >>> index.acos() Float64Index([ 3.141592653589793, 1.1592794807274085, 0.0, 1.5707963267948966, 1.266103672779499], dtype='float64')
-
add(other, fill_value=None, axis=0)¶ Addition of series and other, element-wise (binary operator add).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
all(axis=0, bool_only=None, skipna=True, level=None, **kwargs)¶ Return whether all elements are True in Series.
- Parameters
- skipnabool, default True
Exclude NA/null values. If the entire row/column is NA and skipna is True, then the result will be True, as for an empty row/column. If skipna is False, then NA are treated as True, because these are not equal to zero.
- Returns
- scalar
Notes
Parameters currently not supported are axis, bool_only, level.
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.all() True
-
any(axis=0, bool_only=None, skipna=True, level=None, **kwargs)¶ Return whether any elements is True in Series.
- Parameters
- skipnabool, default True
Exclude NA/null values. If the entire row/column is NA and skipna is True, then the result will be False, as for an empty row/column. If skipna is False, then NA are treated as True, because these are not equal to zero.
- Returns
- scalar
Notes
Parameters currently not supported are axis, bool_only, level.
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.any() True
-
append(to_append, ignore_index=False, verify_integrity=False)¶ Append values from another
Seriesor array-like object. Ifignore_index=True, the index is reset.- Parameters
- to_appendSeries or list/tuple of Series
Series to append with self.
- ignore_indexboolean, default False.
If True, do not use the index.
- verify_integritybool, default False
This Parameter is currently not supported.
- Returns
- Series
A new concatenated series
See also
cudf.core.reshape.concatGeneral function to concatenate DataFrame or Series objects.
Examples
>>> import cudf >>> s1 = cudf.Series([1, 2, 3]) >>> s2 = cudf.Series([4, 5, 6]) >>> s1 0 1 1 2 2 3 dtype: int64 >>> s2 0 4 1 5 2 6 dtype: int64 >>> s1.append(s2) 0 1 1 2 2 3 0 4 1 5 2 6 dtype: int64
>>> s3 = cudf.Series([4, 5, 6], index=[3, 4, 5]) >>> s3 3 4 4 5 5 6 dtype: int64 >>> s1.append(s3) 0 1 1 2 2 3 3 4 4 5 5 6 dtype: int64
With ignore_index set to True:
>>> s1.append(s2, ignore_index=True) 0 1 1 2 2 3 3 4 4 5 5 6 dtype: int64
-
applymap(udf, out_dtype=None)¶ Apply an elementwise function to transform the values in the Column.
The user function is expected to take one argument and return the result, which will be stored to the output Series. The function cannot reference globals except for other simple scalar objects.
- Parameters
- udffunction
Either a callable python function or a python function already decorated by
numba.cuda.jitfor call on the GPU as a device- out_dtypenumpy.dtype; optional
The dtype for use in the output. Only used for
numba.cuda.jitdecorated udf. By default, the result will have the same dtype as the source.
- Returns
- resultSeries
The mask and index are preserved.
Notes
The supported Python features are listed in
with these exceptions:
Math functions in cmath are not supported since libcudf does not have complex number support and output of cmath functions are most likely complex numbers.
These five functions in math are not supported since numba generates multiple PTX functions from them
math.sin()
math.cos()
math.tan()
math.gamma()
math.lgamma()
Series with string dtypes are not supported in applymap method.
Global variables need to be re-defined explicitly inside the udf, as numba considers them to be compile-time constants and there is no known way to obtain value of the global variable.
Examples
Returning a Series of booleans using only a literal pattern.
>>> import cudf >>> s = cudf.Series([1, 10, -10, 200, 100]) >>> s.applymap(lambda x: x) 0 1 1 10 2 -10 3 200 4 100 dtype: int64 >>> s.applymap(lambda x: x in [1, 100, 59]) 0 True 1 False 2 False 3 False 4 True dtype: bool >>> s.applymap(lambda x: x ** 2) 0 1 1 100 2 100 3 40000 4 10000 dtype: int64 >>> s.applymap(lambda x: (x ** 2) + (x / 2)) 0 1.5 1 105.0 2 95.0 3 40100.0 4 10050.0 dtype: float64 >>> def cube_function(a): ... return a ** 3 ... >>> s.applymap(cube_function) 0 1 1 1000 2 -1000 3 8000000 4 1000000 dtype: int64 >>> def custom_udf(x): ... if x > 0: ... return x + 5 ... else: ... return x - 5 ... >>> s.applymap(custom_udf) 0 6 1 15 2 -15 3 205 4 105 dtype: int64
-
argsort(ascending=True, na_position='last')¶ Returns a Series of int64 index that will sort the series.
Uses Thrust sort.
- Returns
- result: Series
-
as_index()¶ Returns a new Series with a RangeIndex.
Examples
>>> s = cudf.Series([1,2,3], index=['a','b','c']) >>> s a 1 b 2 c 3 dtype: int64 >>> s.as_index() 0 1 1 2 2 3 dtype: int64
-
as_mask()¶ Convert booleans to bitmask
- Returns
- device array
-
asin()¶ Get Trigonometric inverse sine, element-wise.
The inverse of sine so that, if y = x.sin(), then x = y.asin()
- Returns
- DataFrame/Series/Index
Result of the trigonometric operation.
Examples
>>> import cudf >>> ser = cudf.Series([-1, 0, 1, 0.32434, 0.5]) >>> ser.asin() 0 -1.570796 1 0.000000 2 1.570796 3 0.330314 4 0.523599 dtype: float64
asin operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-1, 0, 0.5], ... 'second': [0.234, 0.3, 0.1]}) >>> df first second 0 -1.0 0.234 1 0.0 0.300 2 0.5 0.100 >>> df.asin() first second 0 -1.570796 0.236190 1 0.000000 0.304693 2 0.523599 0.100167
asin operation on Index:
>>> index = cudf.Index([-1, 0.4, 1, 0.3]) >>> index Float64Index([-1.0, 0.4, 1.0, 0.3], dtype='float64') >>> index.asin() Float64Index([-1.5707963267948966, 0.41151684606748806, 1.5707963267948966, 0.3046926540153975], dtype='float64')
-
astype(dtype, copy=False, errors='raise')¶ Cast the Series to the given dtype
- Parameters
- dtypedata type, or dict of column name -> data type
Use a numpy.dtype or Python type to cast Series object to the same type. Alternatively, use {col: dtype, …}, where col is a series name and dtype is a numpy.dtype or Python type to cast to.
- copybool, default False
Return a deep-copy when
copy=True. Note by defaultcopy=Falsesetting is used and hence changes to values then may propagate to other cudf objects.- errors{‘raise’, ‘ignore’, ‘warn’}, default ‘raise’
Control raising of exceptions on invalid data for provided dtype. -
raise: allow exceptions to be raised -ignore: suppress exceptions. On error return original object. -warn: prints last exceptions as warnings and return original object.
- Returns
- outSeries
Returns
self.copy(deep=copy)ifdtypeis the same asself.dtype.
-
atan()¶ Get Trigonometric inverse tangent, element-wise.
The inverse of tan so that, if y = x.tan(), then x = y.atan()
- Returns
- DataFrame/Series/Index
Result of the trigonometric operation.
Examples
>>> import cudf >>> ser = cudf.Series([-1, 0, 1, 0.32434, 0.5, -10]) >>> ser 0 -1.00000 1 0.00000 2 1.00000 3 0.32434 4 0.50000 5 -10.00000 dtype: float64 >>> ser.atan() 0 -0.785398 1 0.000000 2 0.785398 3 0.313635 4 0.463648 5 -1.471128 dtype: float64
atan operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-1, -10, 0.5], ... 'second': [0.234, 0.3, 10]}) >>> df first second 0 -1.0 0.234 1 -10.0 0.300 2 0.5 10.000 >>> df.atan() first second 0 -0.785398 0.229864 1 -1.471128 0.291457 2 0.463648 1.471128
atan operation on Index:
>>> index = cudf.Index([-1, 0.4, 1, 0, 0.3]) >>> index Float64Index([-1.0, 0.4, 1.0, 0.0, 0.3], dtype='float64') >>> index.atan() Float64Index([-0.7853981633974483, 0.3805063771123649, 0.7853981633974483, 0.0, 0.2914567944778671], dtype='float64')
-
property
cat¶ Accessor object for categorical properties of the Series values. Be aware that assigning to categories is a inplace operation, while all methods return new categorical data per default.
- Parameters
- columnColumn
- parentSeries or CategoricalIndex
Examples
>>> s = cudf.Series([1,2,3], dtype='category') >>> s >>> s 0 1 1 2 2 3 dtype: category Categories (3, int64): [1, 2, 3] >>> s.cat.categories Int64Index([1, 2, 3], dtype='int64') >>> s.cat.reorder_categories([3,2,1]) 0 1 1 2 2 3 dtype: category Categories (3, int64): [3, 2, 1] >>> s.cat.remove_categories([1]) 0 <NA> 1 2 2 3 dtype: category Categories (2, int64): [2, 3] >>> s.cat.set_categories(list('abcde')) 0 <NA> 1 <NA> 2 <NA> dtype: category Categories (5, object): ['a', 'b', 'c', 'd', 'e'] >>> s.cat.as_ordered() 0 1 1 2 2 3 dtype: category Categories (3, int64): [1 < 2 < 3] >>> s.cat.as_unordered() 0 1 1 2 2 3 dtype: category Categories (3, int64): [1, 2, 3]
-
ceil()¶ Rounds each value upward to the smallest integral value not less than the original.
Returns a new Series.
-
clip(lower=None, upper=None, inplace=False, axis=1)¶ Trim values at input threshold(s).
Assigns values outside boundary to boundary values. Thresholds can be singular values or array like, and in the latter case the clipping is performed element-wise in the specified axis. Currently only axis=1 is supported.
- Parameters
- lowerscalar or array_like, default None
Minimum threshold value. All values below this threshold will be set to it. If it is None, there will be no clipping based on lower. In case of Series/Index, lower is expected to be a scalar or an array of size 1.
- upperscalar or array_like, default None
Maximum threshold value. All values below this threshold will be set to it. If it is None, there will be no clipping based on upper. In case of Series, upper is expected to be a scalar or an array of size 1.
- inplacebool, default False
- Returns
- Clipped DataFrame/Series/Index/MultiIndex
Examples
>>> import cudf >>> df = cudf.DataFrame({"a":[1, 2, 3, 4], "b":['a', 'b', 'c', 'd']}) >>> df.clip(lower=[2, 'b'], upper=[3, 'c']) a b 0 2 b 1 2 b 2 3 c 3 3 c
>>> df.clip(lower=None, upper=[3, 'c']) a b 0 1 a 1 2 b 2 3 c 3 3 c
>>> df.clip(lower=[2, 'b'], upper=None) a b 0 2 b 1 2 b 2 3 c 3 4 d
>>> df.clip(lower=2, upper=3, inplace=True) >>> df a b 0 2 2 1 2 3 2 3 3 3 3 3
>>> import cudf >>> sr = cudf.Series([1, 2, 3, 4]) >>> sr.clip(lower=2, upper=3) 0 2 1 2 2 3 3 3 dtype: int64
>>> sr.clip(lower=None, upper=3) 0 1 1 2 2 3 3 3 dtype: int64
>>> sr.clip(lower=2, upper=None, inplace=True) >>> sr 0 2 1 2 2 3 3 4 dtype: int64
-
copy(deep: bool = True) → T¶ Make a copy of this object’s indices and data.
When
deep=True(default), a new object will be created with a copy of the calling object’s data and indices. Modifications to the data or indices of the copy will not be reflected in the original object (see notes below). Whendeep=False, a new object will be created without copying the calling object’s data or index (only references to the data and index are copied). Any changes to the data of the original will be reflected in the shallow copy (and vice versa).- Parameters
- deepbool, default True
Make a deep copy, including a copy of the data and the indices. With
deep=Falseneither the indices nor the data are copied.
- Returns
- copySeries or DataFrame
Object type matches caller.
- Examples
>>> s = cudf.Series([1, 2], index=["a", "b"]) ..
>>> s ..
- a 1
- b 2
- dtype: int64
>>> s_copy = s.copy() ..
>>> s_copy ..
- a 1
- b 2
- dtype: int64
- Shallow copy versus default (deep) copy:
>>> s = cudf.Series([1, 2], index=["a", "b"]) ..
>>> deep = s.copy() ..
>>> shallow = s.copy(deep=False) ..
- Shallow copy shares data and index with original.
>>> s is shallow ..
- False
>>> s._column is shallow._column and s.index is shallow.index ..
- True
- Deep copy has own copy of data and index.
>>> s is deep ..
- False
>>> s.values is deep.values or s.index is deep.index ..
- False
- Updates to the data shared by shallow copy and original is reflected
- in both; deep copy remains unchanged.
>>> s['a'] = 3 ..
>>> shallow['b'] = 4 ..
>>> s ..
- a 3
- b 4
- dtype: int64
>>> shallow ..
- a 3
- b 4
- dtype: int64
>>> deep ..
- a 1
- b 2
- dtype: int64
-
corr(other, method='pearson', min_periods=None)¶ Calculates the sample correlation between two Series, excluding missing values.
Examples
>>> import cudf >>> ser1 = cudf.Series([0.9, 0.13, 0.62]) >>> ser2 = cudf.Series([0.12, 0.26, 0.51]) >>> ser1.corr(ser2) -0.20454263717316112
-
cos()¶ Get Trigonometric cosine, element-wise.
- Returns
- DataFrame/Series/Index
Result of the trigonometric operation.
Examples
>>> import cudf >>> ser = cudf.Series([0.0, 0.32434, 0.5, 45, 90, 180, 360]) >>> ser 0 0.00000 1 0.32434 2 0.50000 3 45.00000 4 90.00000 5 180.00000 6 360.00000 dtype: float64 >>> ser.cos() 0 1.000000 1 0.947861 2 0.877583 3 0.525322 4 -0.448074 5 -0.598460 6 -0.283691 dtype: float64
cos operation on DataFrame:
>>> df = cudf.DataFrame({'first': [0.0, 5, 10, 15], ... 'second': [100.0, 360, 720, 300]}) >>> df first second 0 0.0 100.0 1 5.0 360.0 2 10.0 720.0 3 15.0 300.0 >>> df.cos() first second 0 1.000000 0.862319 1 0.283662 -0.283691 2 -0.839072 -0.839039 3 -0.759688 -0.022097
cos operation on Index:
>>> index = cudf.Index([-0.4, 100, -180, 90]) >>> index Float64Index([-0.4, 100.0, -180.0, 90.0], dtype='float64') >>> index.cos() Float64Index([ 0.9210609940028851, 0.8623188722876839, -0.5984600690578581, -0.4480736161291701], dtype='float64')
-
count(level=None, **kwargs)¶ Return number of non-NA/null observations in the Series
- Returns
- int
Number of non-null values in the Series.
Notes
Parameters currently not supported is level.
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.count() 5
-
cov(other, min_periods=None)¶ Compute covariance with Series, excluding missing values.
- Parameters
- otherSeries
Series with which to compute the covariance.
- Returns
- float
Covariance between Series and other normalized by N-1 (unbiased estimator).
Notes
min_periods parameter is not yet supported.
Examples
>>> import cudf >>> ser1 = cudf.Series([0.9, 0.13, 0.62]) >>> ser2 = cudf.Series([0.12, 0.26, 0.51]) >>> ser1.cov(ser2) -0.015750000000000004
-
cummax(axis=0, skipna=True, *args, **kwargs)¶ Return cumulative maximum of the Series.
- Parameters
- skipnabool, default True
Exclude NA/null values. If an entire row/column is NA, the result will be NA.
- Returns
- Series
Notes
Parameters currently not supported is axis
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.cummax() 0 1 1 5 2 5 3 5 4 5
-
cummin(axis=None, skipna=True, *args, **kwargs)¶ Return cumulative minimum of the Series.
- Parameters
- skipnabool, default True
Exclude NA/null values. If an entire row/column is NA, the result will be NA.
- Returns
- Series
Notes
Parameters currently not supported is axis
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.cummin() 0 1 1 1 2 1 3 1 4 1
-
cumprod(axis=0, skipna=True, *args, **kwargs)¶ Return cumulative product of the Series.
- Parameters
- skipnabool, default True
Exclude NA/null values. If an entire row/column is NA, the result will be NA.
- Returns
- Series
Notes
Parameters currently not supported is axis
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.cumprod() 0 1 1 5 2 10 3 40 4 120
-
cumsum(axis=0, skipna=True, *args, **kwargs)¶ Return cumulative sum of the Series.
- Parameters
- skipnabool, default True
Exclude NA/null values. If an entire row/column is NA, the result will be NA.
- Returns
- Series
Notes
Parameters currently not supported is axis
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.cumsum() 0 1 1 6 2 8 3 12 4 15
-
property
data¶ The gpu buffer for the data
-
describe(percentiles=None, include=None, exclude=None, datetime_is_numeric=False)¶ Generate descriptive statistics.
Descriptive statistics include those that summarize the central tendency, dispersion and shape of a dataset’s distribution, excluding
NaNvalues.Analyzes both numeric and object series, as well as
DataFramecolumn sets of mixed data types. The output will vary depending on what is provided. Refer to the notes below for more detail.- Parameters
- percentileslist-like of numbers, optional
The percentiles to include in the output. All should fall between 0 and 1. The default is
[.25, .5, .75], which returns the 25th, 50th, and 75th percentiles.- include‘all’, list-like of dtypes or None(default), optional
A list of data types to include in the result. Ignored for
Series. Here are the options:‘all’ : All columns of the input will be included in the output.
A list-like of dtypes : Limits the results to the provided data types. To limit the result to numeric types submit
numpy.number. To limit it instead to object columns submit thenumpy.objectdata type. Strings can also be used in the style ofselect_dtypes(e.g.df.describe(include=['O'])). To select pandas categorical columns, use'category'None (default) : The result will include all numeric columns.
- excludelist-like of dtypes or None (default), optional,
A list of data types to omit from the result. Ignored for
Series. Here are the options:A list-like of dtypes : Excludes the provided data types from the result. To exclude numeric types submit
numpy.number. To exclude object columns submit the data typenumpy.object. Strings can also be used in the style ofselect_dtypes(e.g.df.describe(include=['O'])). To exclude pandas categorical columns, use'category'None (default) : The result will exclude nothing.
- datetime_is_numericbool, default False
For DataFrame input, this also controls whether datetime columns are included by default.
- Returns
- output_frameSeries or DataFrame
Summary statistics of the Series or Dataframe provided.
Notes
For numeric data, the result’s index will include
count,mean,std,min,maxas well as lower,50and upper percentiles. By default the lower percentile is25and the upper percentile is75. The50percentile is the same as the median.For strings dtype or datetime dtype, the result’s index will include
count,unique,top, andfreq. Thetopis the most common value. Thefreqis the most common value’s frequency. Timestamps also include thefirstandlastitems.If multiple object values have the highest count, then the
countandtopresults will be arbitrarily chosen from among those with the highest count.For mixed data types provided via a
DataFrame, the default is to return only an analysis of numeric columns. If the dataframe consists only of object and categorical data without any numeric columns, the default is to return an analysis of both the object and categorical columns. Ifinclude='all'is provided as an option, the result will include a union of attributes of each type.The
includeandexcludeparameters can be used to limit which columns in aDataFrameare analyzed for the output. The parameters are ignored when analyzing aSeries.Examples
Describing a
Seriescontaining numeric values.>>> import cudf >>> s = cudf.Series([1, 2, 3, 4, 5, 6, 7, 8, 9, 10]) >>> s 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 dtype: int64 >>> s.describe() count 10.00000 mean 5.50000 std 3.02765 min 1.00000 25% 3.25000 50% 5.50000 75% 7.75000 max 10.00000 dtype: float64
Describing a categorical
Series.>>> s = cudf.Series(['a', 'b', 'a', 'b', 'c', 'a'], dtype='category') >>> s 0 a 1 b 2 a 3 b 4 c 5 a dtype: category Categories (3, object): ['a', 'b', 'c'] >>> s.describe() count 6 unique 3 top a freq 3 dtype: object
Describing a timestamp
Series.>>> import numpy as np >>> s = cudf.Series([ ... np.datetime64("2000-01-01"), ... np.datetime64("2010-01-01"), ... np.datetime64("2010-01-01") ... ]) >>> s 0 2000-01-01 1 2010-01-01 2 2010-01-01 dtype: datetime64[s] >>> s.describe() count 3 mean 2006-09-01 08:00:00.000000000 min 2000-01-01 00:00:00.000000000 25% 2004-12-31 12:00:00.000000000 50% 2010-01-01 00:00:00.000000000 75% 2010-01-01 00:00:00.000000000 max 2010-01-01 00:00:00.000000000 dtype: object
Describing a
DataFrame. By default only numeric fields are returned.>>> df = cudf.DataFrame({"categorical": cudf.Series(['d', 'e', 'f'], ... dtype='category'), ... "numeric": [1, 2, 3], ... "object": ['a', 'b', 'c'] ... }) >>> df categorical numeric object 0 d 1 a 1 e 2 b 2 f 3 c >>> df.describe() numeric count 3.0 mean 2.0 std 1.0 min 1.0 25% 1.5 50% 2.0 75% 2.5 max 3.0
Describing all columns of a
DataFrameregardless of data type.>>> df.describe(include='all') categorical numeric object count 3 3.0 3 unique 3 <NA> 3 top d <NA> a freq 1 <NA> 1 mean <NA> 2.0 <NA> std <NA> 1.0 <NA> min <NA> 1.0 <NA> 25% <NA> 1.5 <NA> 50% <NA> 2.0 <NA> 75% <NA> 2.5 <NA> max <NA> 3.0 <NA>
Describing a column from a
DataFrameby accessing it as an attribute.>>> df.numeric.describe() count 3.0 mean 2.0 std 1.0 min 1.0 25% 1.5 50% 2.0 75% 2.5 max 3.0 Name: numeric, dtype: float64
Including only numeric columns in a
DataFramedescription.>>> df.describe(include=[np.number]) numeric count 3.0 mean 2.0 std 1.0 min 1.0 25% 1.5 50% 2.0 75% 2.5 max 3.0
Including only string columns in a
DataFramedescription.>>> df.describe(include=[object]) object count 3 unique 3 top a freq 1
Including only categorical columns from a
DataFramedescription.>>> df.describe(include=['category']) categorical count 3 unique 3 top d freq 1
Excluding numeric columns from a
DataFramedescription.>>> df.describe(exclude=[np.number]) categorical object count 3 3 unique 3 3 top d a freq 1 1
Excluding object columns from a
DataFramedescription.>>> df.describe(exclude=[object]) categorical numeric count 3 3.0 unique 3 <NA> top d <NA> freq 1 <NA> mean <NA> 2.0 std <NA> 1.0 min <NA> 1.0 25% <NA> 1.5 50% <NA> 2.0 75% <NA> 2.5 max <NA> 3.0
-
diff(periods=1)¶ Calculate the difference between values at positions i and i - N in an array and store the output in a new array.
Notes
Diff currently only supports float and integer dtype columns with no null values.
-
digitize(bins, right=False)¶ Return the indices of the bins to which each value in series belongs.
- Parameters
- binsnp.array
1-D monotonically, increasing array with same type as this series.
- rightbool
Indicates whether interval contains the right or left bin edge.
- Returns
- A new Series containing the indices.
Notes
Monotonicity of bins is assumed and not checked.
-
drop_duplicates(keep='first', inplace=False, ignore_index=False)¶ Return Series with duplicate values removed
-
dropna(axis=0, inplace=False, how=None)¶ Return a Series with null values removed.
- Parameters
- axis{0 or ‘index’}, default 0
There is only one axis to drop values from.
- inplacebool, default False
If True, do operation inplace and return None.
- howstr, optional
Not in use. Kept for compatibility.
- Returns
- Series
Series with null entries dropped from it.
See also
Series.isnaIndicate null values.
Series.notnaIndicate non-null values.
Series.fillnaReplace null values.
cudf.core.dataframe.DataFrame.dropnaDrop rows or columns which contain null values.
cudf.core.index.Index.dropnaDrop null indices.
Examples
>>> import cudf >>> ser = cudf.Series([1, 2, None]) >>> ser 0 1 1 2 2 null dtype: int64
Drop null values from a Series.
>>> ser.dropna() 0 1 1 2 dtype: int64
Keep the Series with valid entries in the same variable.
>>> ser.dropna(inplace=True) >>> ser 0 1 1 2 dtype: int64
Empty strings are not considered null values. None is considered a null value.
>>> ser = cudf.Series(['', None, 'abc']) >>> ser 0 1 <NA> 2 abc dtype: object >>> ser.dropna() 0 2 abc dtype: object
-
property
dt¶ Accessor object for datetimelike properties of the Series values.
- Returns
- A Series indexed like the original Series.
- Raises
- TypeError if the Series does not contain datetimelike values.
Examples
>>> s.dt.hour >>> s.dt.second >>> s.dt.day
-
property
dtype¶ dtype of the Series
-
property
empty¶ Indicator whether DataFrame or Series is empty.
True if DataFrame/Series is entirely empty (no items), meaning any of the axes are of length 0.
- Returns
- outbool
If DataFrame/Series is empty, return True, if not return False.
Notes
If DataFrame/Series contains only null values, it is still not considered empty. See the example below.
Examples
>>> import cudf >>> df = cudf.DataFrame({'A' : []}) >>> df Empty DataFrame Columns: [A] Index: [] >>> df.empty True
If we only have null values in our DataFrame, it is not considered empty! We will need to drop the null’s to make the DataFrame empty:
>>> df = cudf.DataFrame({'A' : [None, None]}) >>> df A 0 <NA> 1 <NA> >>> df.empty False >>> df.dropna().empty True
Non-empty and empty Series example:
>>> s = cudf.Series([1, 2, None]) >>> s 0 1 1 2 2 <NA> dtype: int64 >>> s.empty False >>> s = cudf.Series([]) >>> s Series([], dtype: float64) >>> s.empty True
-
eq(other, fill_value=None, axis=0)¶ Equal to of series and other, element-wise (binary operator eq).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
equals(other, **kwargs)¶ Test whether two objects contain the same elements. This function allows two Series or DataFrames to be compared against each other to see if they have the same shape and elements. NaNs in the same location are considered equal. The column headers do not need to have the same type.
- Parameters
- otherSeries or DataFrame
The other Series or DataFrame to be compared with the first.
- Returns
- bool
True if all elements are the same in both objects, False otherwise.
Examples
>>> import cudf
Comparing Series with equals:
>>> s = cudf.Series([1, 2, 3]) >>> other = cudf.Series([1, 2, 3]) >>> s.equals(other) True >>> different = cudf.Series([1.5, 2, 3]) >>> s.equals(different) False
Comparing DataFrames with equals:
>>> df = cudf.DataFrame({1: [10], 2: [20]}) >>> df 1 2 0 10 20 >>> exactly_equal = cudf.DataFrame({1: [10], 2: [20]}) >>> exactly_equal 1 2 0 10 20 >>> df.equals(exactly_equal) True
For two DataFrames to compare equal, the types of column values must be equal, but the types of column labels need not:
>>> different_column_type = cudf.DataFrame({1.0: [10], 2.0: [20]}) >>> different_column_type 1.0 2.0 0 10 20 >>> df.equals(different_column_type) True
-
exp()¶ Get the exponential of all elements, element-wise.
Exponential is the inverse of the log function, so that x.exp().log() = x
- Returns
- DataFrame/Series/Index
Result of the element-wise exponential.
Examples
>>> import cudf >>> ser = cudf.Series([-1, 0, 1, 0.32434, 0.5, -10, 100]) >>> ser 0 -1.00000 1 0.00000 2 1.00000 3 0.32434 4 0.50000 5 -10.00000 6 100.00000 dtype: float64 >>> ser.exp() 0 3.678794e-01 1 1.000000e+00 2 2.718282e+00 3 1.383117e+00 4 1.648721e+00 5 4.539993e-05 6 2.688117e+43 dtype: float64
exp operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-1, -10, 0.5], ... 'second': [0.234, 0.3, 10]}) >>> df first second 0 -1.0 0.234 1 -10.0 0.300 2 0.5 10.000 >>> df.exp() first second 0 0.367879 1.263644 1 0.000045 1.349859 2 1.648721 22026.465795
exp operation on Index:
>>> index = cudf.Index([-1, 0.4, 1, 0, 0.3]) >>> index Float64Index([-1.0, 0.4, 1.0, 0.0, 0.3], dtype='float64') >>> index.exp() Float64Index([0.36787944117144233, 1.4918246976412703, 2.718281828459045, 1.0, 1.3498588075760032], dtype='float64')
-
factorize(na_sentinel=- 1)¶ Encode the input values as integer labels
- Parameters
- na_sentinelnumber
Value to indicate missing category.
- Returns
- (labels, cats)(Series, Series)
labels contains the encoded values
cats contains the categories in order that the N-th item corresponds to the (N-1) code.
Examples
>>> import cudf >>> s = cudf.Series(['a', 'a', 'c']) >>> codes, uniques = s.factorize() >>> codes 0 0 1 0 2 1 dtype: int8 >>> uniques 0 a 1 c dtype: object
-
fillna(value=None, method=None, axis=None, inplace=False, limit=None)¶ Fill null values with
valueor specifiedmethod.- Parameters
- valuescalar, Series-like or dict
Value to use to fill nulls. If Series-like, null values are filled with values in corresponding indices. A dict can be used to provide different values to fill nulls in different columns. Cannot be used with
method.- method{‘ffill’, ‘bfill’}, default None
Method to use for filling null values in the dataframe or series. ffill propagates the last non-null values forward to the next non-null value. bfill propagates backward with the next non-null value. Cannot be used with
value.
- Returns
- resultDataFrame
Copy with nulls filled.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, None], 'b': [3, None, 5]}) >>> df a b 0 1 3 1 2 <NA> 2 <NA> 5 >>> df.fillna(4) a b 0 1 3 1 2 4 2 4 5 >>> df.fillna({'a': 3, 'b': 4}) a b 0 1 3 1 2 4 2 3 5
fillnaon a Series object:>>> ser = cudf.Series(['a', 'b', None, 'c']) >>> ser 0 a 1 b 2 <NA> 3 c dtype: object >>> ser.fillna('z') 0 a 1 b 2 z 3 c dtype: object
fillnacan also supports inplace operation:>>> ser.fillna('z', inplace=True) >>> ser 0 a 1 b 2 z 3 c dtype: object >>> df.fillna({'a': 3, 'b': 4}, inplace=True) >>> df a b 0 1 3 1 2 4 2 3 5
fillnaspecified with fillmethod>>> ser = cudf.Series([1, None, None, 2, 3, None, None]) >>> ser.fillna(method='ffill') 0 1 1 1 2 1 3 2 4 3 5 3 6 3 dtype: int64 >>> ser.fillna(method='bfill') 0 1 1 2 2 2 3 2 4 3 5 <NA> 6 <NA> dtype: int64
-
floor()¶ Rounds each value downward to the largest integral value not greater than the original.
Returns a new Series.
-
floordiv(other, fill_value=None, axis=0)¶ Integer division of series and other, element-wise (binary operator floordiv).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
classmethod
from_arrow(array)¶ Convert from PyArrow Array/ChunkedArray to Series.
- Parameters
- arrayPyArrow Array/ChunkedArray
PyArrow Object which has to be converted to cudf Series.
- Returns
- cudf Series
- Raises
- TypeError for invalid input type.
Examples
>>> import cudf >>> import pyarrow as pa >>> cudf.Series.from_arrow(pa.array(["a", "b", None])) 0 a 1 b 2 <NA> dtype: object
-
classmethod
from_categorical(categorical, codes=None)¶ Creates from a pandas.Categorical
If
codesis defined, use it instead ofcategorical.codes
-
classmethod
from_masked_array(data, mask, null_count=None)¶ Create a Series with null-mask. This is equivalent to:
Series(data).set_mask(mask, null_count=null_count)
- Parameters
- data1D array-like
The values. Null values must not be skipped. They can appear as garbage values.
- mask1D array-like
The null-mask. Valid values are marked as
1; otherwise0. The mask bit given the data indexidxis computed as:(mask[idx // 8] >> (idx % 8)) & 1
- null_countint, optional
The number of null values. If None, it is calculated automatically.
-
classmethod
from_pandas(s, nan_as_null=None)¶ Convert from a Pandas Series.
- Parameters
- sPandas Series object
A Pandas Series object which has to be converted to cuDF Series.
- nan_as_nullbool, Default None
If
None/True, convertsnp.nanvalues tonullvalues. IfFalse, leavesnp.nanvalues as is.
- Raises
- TypeError for invalid input type.
Examples
>>> import cudf >>> import pandas as pd >>> import numpy as np >>> data = [10, 20, 30, np.nan] >>> pds = pd.Series(data) >>> cudf.Series.from_pandas(pds) 0 10.0 1 20.0 2 30.0 3 <NA> dtype: float64 >>> cudf.Series.from_pandas(pds, nan_as_null=False) 0 10.0 1 20.0 2 30.0 3 NaN dtype: float64
-
ge(other, fill_value=None, axis=0)¶ Greater than or equal to of series and other, element-wise (binary operator ge).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
groupby(by=None, axis=0, level=None, as_index=True, sort=False, group_keys=True, squeeze=False, observed=False, dropna=True)¶ Group Series using a mapper or by a Series of columns.
A groupby operation involves some combination of splitting the object, applying a function, and combining the results. This can be used to group large amounts of data and compute operations on these groups.
- Parameters
- bymapping, function, label, or list of labels
Used to determine the groups for the groupby. If by is a function, it’s called on each value of the object’s index. If a dict or Series is passed, the Series or dict VALUES will be used to determine the groups (the Series’ values are first aligned; see .align() method). If an cupy array is passed, the values are used as-is determine the groups. A label or list of labels may be passed to group by the columns in self. Notice that a tuple is interpreted as a (single) key.
- levelint, level name, or sequence of such, default None
If the axis is a MultiIndex (hierarchical), group by a particular level or levels.
- as_indexbool, default True
For aggregated output, return object with group labels as the index. Only relevant for DataFrame input. as_index=False is effectively “SQL-style” grouped output.
- sortbool, default False
Sort result by group key. Differ from Pandas, cudf defaults to
Falsefor better performance. Note this does not influence the order of observations within each group. Groupby preserves the order of rows within each group.
- Returns
- SeriesGroupBy
Returns a groupby object that contains information about the groups.
Examples
>>> ser = cudf.Series([390., 350., 30., 20.], ... index=['Falcon', 'Falcon', 'Parrot', 'Parrot'], ... name="Max Speed") >>> ser Falcon 390.0 Falcon 350.0 Parrot 30.0 Parrot 20.0 Name: Max Speed, dtype: float64 >>> ser.groupby(level=0).mean() Falcon 370.0 Parrot 25.0 Name: Max Speed, dtype: float64 >>> ser.groupby(ser > 100).mean() Max Speed False 25.0 True 370.0 Name: Max Speed, dtype: float64
-
gt(other, fill_value=None, axis=0)¶ Greater than of series and other, element-wise (binary operator gt).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
property
has_nulls¶ Indicator whether Series contains null values.
- Returns
- outbool
If Series has atleast one null value, return True, if not return False.
-
hash_encode(stop, use_name=False)¶ Encode column values as ints in [0, stop) using hash function.
- Parameters
- stopint
The upper bound on the encoding range.
- use_namebool
If
Truethen combine hashed column values with hashed column name. This is useful for when the same values in different columns should be encoded with different hashed values.
- Returns
- resultSeries
The encoded Series.
-
hash_values()¶ Compute the hash of values in this column.
-
head(n=5)¶ Return the first n rows. This function returns the first n rows for the object based on position. It is useful for quickly testing if your object has the right type of data in it. For negative values of n, this function returns all rows except the last n rows, equivalent to
df[:-n].- Parameters
- nint, default 5
Number of rows to select.
- Returns
- same type as caller
The first n rows of the caller object.
See also
Series.tailReturns the last n rows.
Examples
>>> ser = cudf.Series(['alligator', 'bee', 'falcon', ... 'lion', 'monkey', 'parrot', 'shark', 'whale', 'zebra']) >>> ser 0 alligator 1 bee 2 falcon 3 lion 4 monkey 5 parrot 6 shark 7 whale 8 zebra dtype: object
Viewing the first 5 lines
>>> ser.head() 0 alligator 1 bee 2 falcon 3 lion 4 monkey dtype: object
Viewing the first n lines (three in this case)
>>> ser.head(3) 0 alligator 1 bee 2 falcon dtype: object
For negative values of n
>>> ser.head(-3) 0 alligator 1 bee 2 falcon 3 lion 4 monkey 5 parrot dtype: object
-
property
iloc¶ Select values by position.
See also
-
property
index¶ The index object
-
interleave_columns()¶ Interleave Series columns of a table into a single column.
Converts the column major table cols into a row major column.
- Parameters
- colsinput Table containing columns to interleave.
- Returns
- The interleaved columns as a single column
Examples
>>> df = DataFrame([['A1', 'A2', 'A3'], ['B1', 'B2', 'B3']]) >>> df 0 [A1, A2, A3] 1 [B1, B2, B3] >>> df.interleave_columns() 0 A1 1 B1 2 A2 3 B2 4 A3 5 B3
-
property
is_monotonic¶ Return boolean if values in the object are monotonic_increasing.
- Returns
- outbool
-
property
is_monotonic_decreasing¶ Return boolean if values in the object are monotonic_decreasing.
- Returns
- outbool
-
property
is_monotonic_increasing¶ Return boolean if values in the object are monotonic_increasing.
- Returns
- outbool
-
property
is_unique¶ Return boolean if values in the object are unique.
- Returns
- outbool
-
isin(values)¶ Check whether values are contained in Series.
- Parameters
- valuesset or list-like
The sequence of values to test. Passing in a single string will raise a TypeError. Instead, turn a single string into a list of one element.
- Returns
- resultSeries
Series of booleans indicating if each element is in values.
- Raises
- TypeError
If values is a string
-
isna()¶ Identify missing values.
Return a boolean same-sized object indicating if the values are
<NA>.<NA>values gets mapped toTruevalues. Everything else gets mapped toFalsevalues.<NA>values include:Values where null mask is set.
NaNin float dtype.NaTin datetime64 and timedelta64 types.
Characters such as empty strings
''orinfincase of float are not considered<NA>values.- Returns
- DataFrame/Series/Index
Mask of bool values for each element in the object that indicates whether an element is an NA value.
Examples
Show which entries in a DataFrame are NA.
>>> import cudf >>> import numpy as np >>> import pandas as pd >>> df = cudf.DataFrame({'age': [5, 6, np.NaN], ... 'born': [pd.NaT, pd.Timestamp('1939-05-27'), ... pd.Timestamp('1940-04-25')], ... 'name': ['Alfred', 'Batman', ''], ... 'toy': [None, 'Batmobile', 'Joker']}) >>> df age born name toy 0 5 <NA> Alfred <NA> 1 6 1939-05-27 00:00:00.000000 Batman Batmobile 2 <NA> 1940-04-25 00:00:00.000000 Joker >>> df.isnull() age born name toy 0 False True False True 1 False False False False 2 True False False False
Show which entries in a Series are NA.
>>> ser = cudf.Series([5, 6, np.NaN, np.inf, -np.inf]) >>> ser 0 5.0 1 6.0 2 <NA> 3 Inf 4 -Inf dtype: float64 >>> ser.isnull() 0 False 1 False 2 True 3 False 4 False dtype: bool
Show which entries in an Index are NA.
>>> idx = cudf.Index([1, 2, None, np.NaN, 0.32, np.inf]) >>> idx Float64Index([1.0, 2.0, <NA>, <NA>, 0.32, Inf], dtype='float64') >>> idx.isnull() GenericIndex([False, False, True, True, False, False], dtype='bool')
-
isnull()¶ Identify missing values.
Return a boolean same-sized object indicating if the values are
<NA>.<NA>values gets mapped toTruevalues. Everything else gets mapped toFalsevalues.<NA>values include:Values where null mask is set.
NaNin float dtype.NaTin datetime64 and timedelta64 types.
Characters such as empty strings
''orinfincase of float are not considered<NA>values.- Returns
- DataFrame/Series/Index
Mask of bool values for each element in the object that indicates whether an element is an NA value.
Examples
Show which entries in a DataFrame are NA.
>>> import cudf >>> import numpy as np >>> import pandas as pd >>> df = cudf.DataFrame({'age': [5, 6, np.NaN], ... 'born': [pd.NaT, pd.Timestamp('1939-05-27'), ... pd.Timestamp('1940-04-25')], ... 'name': ['Alfred', 'Batman', ''], ... 'toy': [None, 'Batmobile', 'Joker']}) >>> df age born name toy 0 5 <NA> Alfred <NA> 1 6 1939-05-27 00:00:00.000000 Batman Batmobile 2 <NA> 1940-04-25 00:00:00.000000 Joker >>> df.isnull() age born name toy 0 False True False True 1 False False False False 2 True False False False
Show which entries in a Series are NA.
>>> ser = cudf.Series([5, 6, np.NaN, np.inf, -np.inf]) >>> ser 0 5.0 1 6.0 2 <NA> 3 Inf 4 -Inf dtype: float64 >>> ser.isnull() 0 False 1 False 2 True 3 False 4 False dtype: bool
Show which entries in an Index are NA.
>>> idx = cudf.Index([1, 2, None, np.NaN, 0.32, np.inf]) >>> idx Float64Index([1.0, 2.0, <NA>, <NA>, 0.32, Inf], dtype='float64') >>> idx.isnull() GenericIndex([False, False, True, True, False, False], dtype='bool')
-
keys()¶ Return alias for index.
- Returns
- Index
Index of the Series.
Examples
>>> import cudf >>> sr = cudf.Series([10, 11, 12, 13, 14, 15]) >>> sr 0 10 1 11 2 12 3 13 4 14 5 15 dtype: int64
>>> sr.keys() RangeIndex(start=0, stop=6) >>> sr = cudf.Series(['a', 'b', 'c']) >>> sr 0 a 1 b 2 c dtype: object >>> sr.keys() RangeIndex(start=0, stop=3) >>> sr = cudf.Series([1, 2, 3], index=['a', 'b', 'c']) >>> sr a 1 b 2 c 3 dtype: int64 >>> sr.keys() StringIndex(['a' 'b' 'c'], dtype='object')
-
kurt(axis=None, skipna=None, level=None, numeric_only=None, **kwargs)¶ Return Fisher’s unbiased kurtosis of a sample.
Kurtosis obtained using Fisher’s definition of kurtosis (kurtosis of normal == 0.0). Normalized by N-1.
- Parameters
- skipnabool, default True
Exclude NA/null values when computing the result.
- Returns
- scalar
Notes
Parameters currently not supported are axis, level and numeric_only
-
kurtosis(axis=None, skipna=None, level=None, numeric_only=None, **kwargs)¶ Return Fisher’s unbiased kurtosis of a sample.
Kurtosis obtained using Fisher’s definition of kurtosis (kurtosis of normal == 0.0). Normalized by N-1.
- Parameters
- skipnabool, default True
Exclude NA/null values when computing the result.
- Returns
- scalar
Notes
Parameters currently not supported are axis, level and numeric_only
-
label_encoding(cats, dtype=None, na_sentinel=- 1)¶ Perform label encoding
- Parameters
- valuessequence of input values
- dtypenumpy.dtype; optional
Specifies the output dtype. If None is given, the smallest possible integer dtype (starting with np.int8) is used.
- na_sentinelnumber, default -1
Value to indicate missing category.
- Returns
- A sequence of encoded labels with value between 0 and n-1 classes(cats)
Examples
>>> import cudf >>> s = cudf.Series([1, 2, 3, 4, 10]) >>> s.label_encoding([2, 3]) 0 -1 1 0 2 1 3 -1 4 -1 dtype: int8
na_sentinel parameter can be used to control the value when there is no encoding.
>>> s.label_encoding([2, 3], na_sentinel=10) 0 10 1 0 2 1 3 10 4 10 dtype: int8
When none of cats values exist in s, entire Series will be na_sentinel.
>>> s.label_encoding(['a', 'b', 'c']) 0 -1 1 -1 2 -1 3 -1 4 -1 dtype: int8
-
le(other, fill_value=None, axis=0)¶ Less than or equal to of series and other, element-wise (binary operator le).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
property
loc¶ Select values by label.
See also
-
log()¶ Get the natural logarithm of all elements, element-wise.
Natural logarithm is the inverse of the exp function, so that x.log().exp() = x
- Returns
- DataFrame/Series/Index
Result of the element-wise natural logarithm.
Examples
>>> import cudf >>> ser = cudf.Series([-1, 0, 1, 0.32434, 0.5, -10, 100]) >>> ser 0 -1.00000 1 0.00000 2 1.00000 3 0.32434 4 0.50000 5 -10.00000 6 100.00000 dtype: float64 >>> ser.log() 0 NaN 1 -inf 2 0.000000 3 -1.125963 4 -0.693147 5 NaN 6 4.605170 dtype: float64
log operation on DataFrame:
>>> df = cudf.DataFrame({'first': [-1, -10, 0.5], ... 'second': [0.234, 0.3, 10]}) >>> df first second 0 -1.0 0.234 1 -10.0 0.300 2 0.5 10.000 >>> df.log() first second 0 NaN -1.452434 1 NaN -1.203973 2 -0.693147 2.302585
log operation on Index:
>>> index = cudf.Index([10, 11, 500.0]) >>> index Float64Index([10.0, 11.0, 500.0], dtype='float64') >>> index.log() Float64Index([2.302585092994046, 2.3978952727983707, 6.214608098422191], dtype='float64')
-
lt(other, fill_value=None, axis=0)¶ Less than of series and other, element-wise (binary operator lt).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
map(arg, na_action=None) → Series¶ Map values of Series according to input correspondence.
Used for substituting each value in a Series with another value, that may be derived from a function, a
dictor aSeries.- Parameters
- argfunction, collections.abc.Mapping subclass or Series
Mapping correspondence.
- na_action{None, ‘ignore’}, default None
If ‘ignore’, propagate NaN values, without passing them to the mapping correspondence.
- Returns
- Series
Same index as caller.
Notes
Please note map currently only supports fixed-width numeric type functions.
Examples
>>> s = cudf.Series(['cat', 'dog', np.nan, 'rabbit']) >>> s 0 cat 1 dog 2 <NA> 3 rabbit dtype: object
mapaccepts adictor aSeries. Values that are not found in thedictare converted toNaN, default values in dicts are currently not supported.:>>> s.map({'cat': 'kitten', 'dog': 'puppy'}) 0 kitten 1 puppy 2 <NA> 3 <NA> dtype: object
It also accepts numeric functions:
>>> s = cudf.Series([1, 2, 3, 4, np.nan]) >>> s.map(lambda x: x ** 2) 0 1 1 4 2 9 3 16 4 <NA> dtype: int64
-
mask(cond, other=None, inplace=False)¶ Replace values where the condition is True.
- Parameters
- condbool Series/DataFrame, array-like
Where cond is False, keep the original value. Where True, replace with corresponding value from other. Callables are not supported.
- other: scalar, list of scalars, Series/DataFrame
Entries where cond is True are replaced with corresponding value from other. Callables are not supported. Default is None.
DataFrame expects only Scalar or array like with scalars or dataframe with same dimension as self.
Series expects only scalar or series like with same length
- inplacebool, default False
Whether to perform the operation in place on the data.
- Returns
- Same type as caller
Examples
>>> import cudf >>> df = cudf.DataFrame({"A":[1, 4, 5], "B":[3, 5, 8]}) >>> df.mask(df % 2 == 0, [-1, -1]) A B 0 1 3 1 -1 5 2 5 -1
>>> ser = cudf.Series([4, 3, 2, 1, 0]) >>> ser.mask(ser > 2, 10) 0 10 1 10 2 2 3 1 4 0 dtype: int64 >>> ser.mask(ser > 2) 0 <NA> 1 <NA> 2 2 3 1 4 0 dtype: int64
-
max(axis=None, skipna=None, dtype=None, level=None, numeric_only=None, **kwargs)¶ Return the maximum of the values in the Series.
- Parameters
- skipnabool, default True
Exclude NA/null values when computing the result.
- dtypedata type
Data type to cast the result to.
- Returns
- scalar
Notes
Parameters currently not supported are axis, level, numeric_only.
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.max() 5
-
mean(axis=None, skipna=None, level=None, numeric_only=None, **kwargs)¶ Return the mean of the values in the series.
- Parameters
- skipnabool, default True
Exclude NA/null values when computing the result.
- Returns
- scalar
Notes
Parameters currently not supported are axis, level and numeric_only
Examples
>>> import cudf >>> ser = cudf.Series([10, 25, 3, 25, 24, 6]) >>> ser.mean() 15.5
-
median(axis=None, skipna=None, level=None, numeric_only=None, **kwargs)¶ Return the median of the values for the requested axis.
- Parameters
- skipnabool, default True
Exclude NA/null values when computing the result.
- Returns
- scalar
Notes
Parameters currently not supported are axis, level and numeric_only
Examples
>>> import cudf >>> ser = cudf.Series([10, 25, 3, 25, 24, 6]) >>> ser 0 10 1 25 2 3 3 25 4 24 5 6 dtype: int64 >>> ser.median() 17.0
-
memory_usage(index=True, deep=False)¶ Return the memory usage of the Series.
The memory usage can optionally include the contribution of the index and of elements of object dtype.
- Parameters
- indexbool, default True
Specifies whether to include the memory usage of the Series index.
- deepbool, default False
If True, introspect the data deeply by interrogating object dtypes for system-level memory consumption, and include it in the returned value.
- Returns
- int
Bytes of memory consumed.
See also
cudf.core.dataframe.DataFrame.memory_usageBytes consumed by a DataFrame.
Examples
>>> s = cudf.Series(range(3), index=['a','b','c']) >>> s.memory_usage() 48
Not including the index gives the size of the rest of the data, which is necessarily smaller:
>>> s.memory_usage(index=False) 24
-
min(axis=None, skipna=None, dtype=None, level=None, numeric_only=None, **kwargs)¶ Return the minimum of the values in the Series.
- Parameters
- skipnabool, default True
Exclude NA/null values when computing the result.
- dtypedata type
Data type to cast the result to.
- Returns
- scalar
Notes
Parameters currently not supported are axis, level, numeric_only.
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.min() 1
-
mod(other, fill_value=None, axis=0)¶ Modulo of series and other, element-wise (binary operator mod).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
mode(dropna=True)¶ Return the mode(s) of the dataset.
Always returns Series even if only one value is returned.
- Parameters
- dropnabool, default True
Don’t consider counts of NA/NaN/NaT.
- Returns
- Series
Modes of the Series in sorted order.
Examples
>>> import cudf >>> series = cudf.Series([7, 6, 5, 4, 3, 2, 1]) >>> series 0 7 1 6 2 5 3 4 4 3 5 2 6 1 dtype: int64 >>> series.mode() 0 1 1 2 2 3 3 4 4 5 5 6 6 7 dtype: int64
We can include
<NA>values in mode by passingdropna=False.>>> series = cudf.Series([7, 4, 3, 3, 7, None, None]) >>> series 0 7 1 4 2 3 3 3 4 7 5 <NA> 6 <NA> dtype: int64 >>> series.mode() 0 3 1 7 dtype: int64 >>> series.mode(dropna=False) 0 3 1 7 2 <NA> dtype: int64
-
mul(other, fill_value=None, axis=0)¶ Multiplication of series and other, element-wise (binary operator mul).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
property
name¶ Returns name of the Series.
-
nans_to_nulls()¶ Convert nans (if any) to nulls
-
property
ndim¶ Dimension of the data. Series ndim is always 1.
-
ne(other, fill_value=None, axis=0)¶ Not equal to of series and other, element-wise (binary operator ne).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
nlargest(n=5, keep='first')¶ Returns a new Series of the n largest element.
-
notna()¶ Identify non-missing values.
Return a boolean same-sized object indicating if the values are not
<NA>. Non-missing values get mapped toTrue.<NA>values get mapped toFalsevalues.<NA>values include:Values where null mask is set.
NaNin float dtype.NaTin datetime64 and timedelta64 types.
Characters such as empty strings
''orinfincase of float are not considered<NA>values.- Returns
- DataFrame/Series/Index
Mask of bool values for each element in the object that indicates whether an element is not an NA value.
Examples
Show which entries in a DataFrame are NA.
>>> import cudf >>> import numpy as np >>> import pandas as pd >>> df = cudf.DataFrame({'age': [5, 6, np.NaN], ... 'born': [pd.NaT, pd.Timestamp('1939-05-27'), ... pd.Timestamp('1940-04-25')], ... 'name': ['Alfred', 'Batman', ''], ... 'toy': [None, 'Batmobile', 'Joker']}) >>> df age born name toy 0 5 <NA> Alfred <NA> 1 6 1939-05-27 00:00:00.000000 Batman Batmobile 2 <NA> 1940-04-25 00:00:00.000000 Joker >>> df.notnull() age born name toy 0 True False True False 1 True True True True 2 False True True True
Show which entries in a Series are NA.
>>> ser = cudf.Series([5, 6, np.NaN, np.inf, -np.inf]) >>> ser 0 5.0 1 6.0 2 <NA> 3 Inf 4 -Inf dtype: float64 >>> ser.notnull() 0 True 1 True 2 False 3 True 4 True dtype: bool
Show which entries in an Index are NA.
>>> idx = cudf.Index([1, 2, None, np.NaN, 0.32, np.inf]) >>> idx Float64Index([1.0, 2.0, <NA>, <NA>, 0.32, Inf], dtype='float64') >>> idx.notnull() GenericIndex([True, True, False, False, True, True], dtype='bool')
-
notnull()¶ Identify non-missing values.
Return a boolean same-sized object indicating if the values are not
<NA>. Non-missing values get mapped toTrue.<NA>values get mapped toFalsevalues.<NA>values include:Values where null mask is set.
NaNin float dtype.NaTin datetime64 and timedelta64 types.
Characters such as empty strings
''orinfincase of float are not considered<NA>values.- Returns
- DataFrame/Series/Index
Mask of bool values for each element in the object that indicates whether an element is not an NA value.
Examples
Show which entries in a DataFrame are NA.
>>> import cudf >>> import numpy as np >>> import pandas as pd >>> df = cudf.DataFrame({'age': [5, 6, np.NaN], ... 'born': [pd.NaT, pd.Timestamp('1939-05-27'), ... pd.Timestamp('1940-04-25')], ... 'name': ['Alfred', 'Batman', ''], ... 'toy': [None, 'Batmobile', 'Joker']}) >>> df age born name toy 0 5 <NA> Alfred <NA> 1 6 1939-05-27 00:00:00.000000 Batman Batmobile 2 <NA> 1940-04-25 00:00:00.000000 Joker >>> df.notnull() age born name toy 0 True False True False 1 True True True True 2 False True True True
Show which entries in a Series are NA.
>>> ser = cudf.Series([5, 6, np.NaN, np.inf, -np.inf]) >>> ser 0 5.0 1 6.0 2 <NA> 3 Inf 4 -Inf dtype: float64 >>> ser.notnull() 0 True 1 True 2 False 3 True 4 True dtype: bool
Show which entries in an Index are NA.
>>> idx = cudf.Index([1, 2, None, np.NaN, 0.32, np.inf]) >>> idx Float64Index([1.0, 2.0, <NA>, <NA>, 0.32, Inf], dtype='float64') >>> idx.notnull() GenericIndex([True, True, False, False, True, True], dtype='bool')
-
nsmallest(n=5, keep='first')¶ Returns a new Series of the n smallest element.
-
property
null_count¶ Number of null values
-
property
nullable¶ A boolean indicating whether a null-mask is needed
-
property
nullmask¶ The gpu buffer for the null-mask
-
nunique(method='sort', dropna=True)¶ Returns the number of unique values of the Series: approximate version, and exact version to be moved to libgdf
-
one_hot_encoding(cats, dtype='float64')¶ Perform one-hot-encoding
- Parameters
- catssequence of values
values representing each category.
- dtypenumpy.dtype
specifies the output dtype.
- Returns
- Sequence
A sequence of new series for each category. Its length is determined by the length of
cats.
-
pipe(func, *args, **kwargs)¶ Apply
func(self, *args, **kwargs).- Parameters
- funcfunction
Function to apply to the Series/DataFrame/Index.
args, andkwargsare passed intofunc. Alternatively a(callable, data_keyword)tuple wheredata_keywordis a string indicating the keyword ofcallablethat expects the Series/DataFrame/Index.- argsiterable, optional
Positional arguments passed into
func.- kwargsmapping, optional
A dictionary of keyword arguments passed into
func.
- Returns
- objectthe return type of
func.
- objectthe return type of
Examples
Use
.pipewhen chaining together functions that expect Series, DataFrames or GroupBy objects. Instead of writing>>> func(g(h(df), arg1=a), arg2=b, arg3=c)
You can write
>>> (df.pipe(h) ... .pipe(g, arg1=a) ... .pipe(func, arg2=b, arg3=c) ... )
If you have a function that takes the data as (say) the second argument, pass a tuple indicating which keyword expects the data. For example, suppose
ftakes its data asarg2:>>> (df.pipe(h) ... .pipe(g, arg1=a) ... .pipe((func, 'arg2'), arg1=a, arg3=c) ... )
-
pow(other, fill_value=None, axis=0)¶ Exponential power of series and other, element-wise (binary operator pow).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
prod(axis=None, skipna=None, dtype=None, level=None, numeric_only=None, min_count=0, **kwargs)¶ Return product of the values in the series
- Parameters
- skipnabool, default True
Exclude NA/null values when computing the result.
- dtypedata type
Data type to cast the result to.
- min_countint, default 0
The required number of valid values to perform the operation. If fewer than min_count non-NA values are present the result will be NA.
The default being 0. This means the sum of an all-NA or empty Series is 0, and the product of an all-NA or empty Series is 1.
- Returns
- scalar
Notes
Parameters currently not supported are axis, level, numeric_only.
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.prod() 120
-
product(axis=None, skipna=None, dtype=None, level=None, numeric_only=None, min_count=0, **kwargs)¶ Return product of the values in the Series.
- Parameters
- skipnabool, default True
Exclude NA/null values when computing the result.
- dtypedata type
Data type to cast the result to.
- min_countint, default 0
The required number of valid values to perform the operation. If fewer than min_count non-NA values are present the result will be NA.
The default being 0. This means the sum of an all-NA or empty Series is 0, and the product of an all-NA or empty Series is 1.
- Returns
- scalar
Notes
Parameters currently not supported are axis, level, numeric_only.
Examples
>>> import cudf >>> ser = cudf.Series([1, 5, 2, 4, 3]) >>> ser.product() 120
-
quantile(q=0.5, interpolation='linear', exact=True, quant_index=True)¶ Return values at the given quantile.
- Parameters
- qfloat or array-like, default 0.5 (50% quantile)
0 <= q <= 1, the quantile(s) to compute
- interpolation{’linear’, ‘lower’, ‘higher’, ‘midpoint’, ‘nearest’}
This optional parameter specifies the interpolation method to use, when the desired quantile lies between two data points i and j:
- columnslist of str
List of column names to include.
- exactboolean
Whether to use approximate or exact quantile algorithm.
- quant_indexboolean
Whether to use the list of quantiles as index.
- Returns
- float or Series
If
qis an array, a Series will be returned where the index isqand the values are the quantiles, otherwise a float will be returned.
-
radd(other, fill_value=None, axis=0)¶ Addition of series and other, element-wise (binary operator radd).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
rank(axis=0, method='average', numeric_only=None, na_option='keep', ascending=True, pct=False)¶ Compute numerical data ranks (1 through n) along axis. By default, equal values are assigned a rank that is the average of the ranks of those values.
- Parameters
- axis{0 or ‘index’, 1 or ‘columns’}, default 0
Index to direct ranking.
- method{‘average’, ‘min’, ‘max’, ‘first’, ‘dense’}, default ‘average’
How to rank the group of records that have the same value (i.e. ties): * average: average rank of the group * min: lowest rank in the group * max: highest rank in the group * first: ranks assigned in order they appear in the array * dense: like ‘min’, but rank always increases by 1 between groups.
- numeric_onlybool, optional
For DataFrame objects, rank only numeric columns if set to True.
- na_option{‘keep’, ‘top’, ‘bottom’}, default ‘keep’
How to rank NaN values: * keep: assign NaN rank to NaN values * top: assign smallest rank to NaN values if ascending * bottom: assign highest rank to NaN values if ascending.
- ascendingbool, default True
Whether or not the elements should be ranked in ascending order.
- pctbool, default False
Whether or not to display the returned rankings in percentile form.
- Returns
- same type as caller
Return a Series or DataFrame with data ranks as values.
-
reindex(index=None, copy=True)¶ Return a Series that conforms to a new index
- Parameters
- indexIndex, Series-convertible, default None
- copyboolean, default True
- Returns
- A new Series that conforms to the supplied index
-
rename(index=None, copy=True)¶ Alter Series name
Change Series.name with a scalar value
- Parameters
- indexScalar, optional
Scalar to alter the Series.name attribute
- copyboolean, default True
Also copy underlying data
- Returns
- Series
Notes
- Difference from pandas:
Supports scalar values only for changing name attribute
Not supporting : inplace, level
-
repeat(repeats, axis=None)¶ Repeats elements consecutively.
Returns a new object of caller type(DataFrame/Series/Index) where each element of the current object is repeated consecutively a given number of times.
- Parameters
- repeatsint, or array of ints
The number of repetitions for each element. This should be a non-negative integer. Repeating 0 times will return an empty object.
- Returns
- Series/DataFrame/Index
A newly created object of same type as caller with repeated elements.
Examples
>>> import cudf >>> df = cudf.DataFrame({'a': [1, 2, 3], 'b': [10, 20, 30]}) >>> df a b 0 1 10 1 2 20 2 3 30 >>> df.repeat(3) a b 0 1 10 0 1 10 0 1 10 1 2 20 1 2 20 1 2 20 2 3 30 2 3 30 2 3 30
Repeat on Series
>>> s = cudf.Series([0, 2]) >>> s 0 0 1 2 dtype: int64 >>> s.repeat([3, 4]) 0 0 0 0 0 0 1 2 1 2 1 2 1 2 dtype: int64 >>> s.repeat(2) 0 0 0 0 1 2 1 2 dtype: int64
Repeat on Index
>>> index = cudf.Index([10, 22, 33, 55]) >>> index Int64Index([10, 22, 33, 55], dtype='int64') >>> index.repeat(5) Int64Index([10, 10, 10, 10, 10, 22, 22, 22, 22, 22, 33, 33, 33, 33, 33, 55, 55, 55, 55, 55], dtype='int64')
-
replace(to_replace=None, value=None, inplace=False, limit=None, regex=False, method=None)¶ Replace values given in
to_replacewithvalue.- Parameters
- to_replacenumeric, str or list-like
Value(s) to replace.
- numeric or str:
values equal to
to_replacewill be replaced withvalue
- list of numeric or str:
If
valueis also list-like,to_replaceandvaluemust be of same length.
- dict:
Dicts can be used to specify different replacement values for different existing values. For example, {‘a’: ‘b’, ‘y’: ‘z’} replaces the value ‘a’ with ‘b’ and ‘y’ with ‘z’. To use a dict in this way the
valueparameter should beNone.
- valuescalar, dict, list-like, str, default None
Value to replace any values matching
to_replacewith.- inplacebool, default False
If True, in place.
- Returns
- resultSeries
Series after replacement. The mask and index are preserved.
- Raises
- TypeError
If
to_replaceis not a scalar, array-like, dict, or NoneIf
to_replaceis a dict and value is not a list, dict, or Series
- ValueError
If a list is passed to
to_replaceandvaluebut they are not the same length.
See also
Notes
Parameters that are currently not supported are: limit, regex, method
Examples
Scalar
to_replaceandvalue>>> import cudf >>> s = cudf.Series([0, 1, 2, 3, 4]) >>> s 0 0 1 1 2 2 3 3 4 4 dtype: int64 >>> s.replace(0, 5) 0 5 1 1 2 2 3 3 4 4 dtype: int64
List-like
to_replace>>> s.replace([1, 2], 10) 0 0 1 10 2 10 3 3 4 4 dtype: int64
dict-like
to_replace>>> s.replace({1:5, 3:50}) 0 0 1 5 2 2 3 50 4 4 dtype: int64 >>> s = cudf.Series(['b', 'a', 'a', 'b', 'a']) >>> s 0 b 1 a 2 a 3 b 4 a dtype: object >>> s.replace({'a': None}) 0 b 1 <NA> 2 <NA> 3 b 4 <NA> dtype: object
If there is a mimatch in types of the values in
to_replace&valuewith the actual series, then cudf exhibits different behaviour with respect to pandas and the pairs are ignored silently:>>> s = cudf.Series(['b', 'a', 'a', 'b', 'a']) >>> s 0 b 1 a 2 a 3 b 4 a dtype: object >>> s.replace('a', 1) 0 b 1 a 2 a 3 b 4 a dtype: object >>> s.replace(['a', 'c'], [1, 2]) 0 b 1 a 2 a 3 b 4 a dtype: object
-
reset_index(drop=False, inplace=False)¶ Reset index to RangeIndex
-
reverse()¶ Reverse the Series
-
rfloordiv(other, fill_value=None, axis=0)¶ Integer division of series and other, element-wise (binary operator rfloordiv).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
- Returns
- Series
Result of the arithmetic operation.
Examples
>>> import cudf >>> s = cudf.Series([1, 2, 10, 17]) >>> s 0 1 1 2 2 10 3 17 dtype: int64 >>> s.rfloordiv(100) 0 100 1 50 2 10 3 5 dtype: int64 >>> s = cudf.Series([10, 20, None]) >>> s 0 10 1 20 2 <NA> dtype: int64 >>> s.rfloordiv(200) 0 20 1 10 2 <NA> dtype: int64 >>> s.rfloordiv(200, fill_value=2) 0 20 1 10 2 100 dtype: int64
-
rmod(other, fill_value=None, axis=0)¶ Modulo of series and other, element-wise (binary operator rmod).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
rmul(other, fill_value=None, axis=0)¶ Multiplication of series and other, element-wise (binary operator rmul).
- Parameters
- otherSeries or scalar value
- fill_valueNone or value
Value to fill nulls with before computation. If data in both corresponding Series locations is null the result will be null
-
rolling(window, min_periods=None, center=False, axis=0, win_type=None)¶ Rolling window calculations.
- Parameters
- windowint or offset
Size of the window, i.e., the number of observations used to calculate the statistic. For datetime indexes, an offset can be provided instead of an int. The offset must be convertible to a timedelta. As opposed to a fixed window size, each window will be sized to accommodate observations within the time period specified by the offset.
- min_periodsint, optional
The minimum number of observations in the window that are required to be non-null, so that the result is non-null. If not provided or
None,min_periodsis equal to the window size.- centerbool, optional
If
True, the result is set at the center of the window. If