ai.rapids.cudf

Class Schema

java.lang.Object

ai.rapids.cudf.Schema

public class Schema
extends Object

The schema of data to be read in.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	Schema.Builder

Field Summary

Fields

Modifier and Type	Field and Description
static Schema	INFERRED

Method Summary

Modifier and Type	Method and Description
HostColumnVector.DataType	asHostDataType()
static Schema.Builder	builder()
Schema	getChild(int i) Get the schema of a child element.
DType[]	getChildTypes() Get types of the top level child columns in schema.
String[]	getColumnNames() Get names of the top level child columns in schema.
String[]	getFlattenedColumnNames() Get names of the columns flattened from all levels in schema by depth-first traversal.
int[]	getFlattenedDecimalPrecisions() Get decimal precisions of the columns' types flattened from all levels in schema by depth-first traversal.
int[]	getFlattenedNumChildren() Get numbers of child columns for each level in schema.
int[]	getFlattenedTypeIds() Get type ids of the columns flattened from all levels in schema by depth-first traversal.
DType[]	getFlattenedTypes() Get the types of the columns in schema flattened from all levels by depth-first traversal.
int[]	getFlattenedTypeScales() Get scales of the columns' types flattened from all levels in schema by depth-first traversal.
int	getNumChildren() Get number of top level child columns in schema.
DType	getType()
boolean	hasNestedChildren() This is really for a top level struct schema where it is nested, but for things like CSV we care that it does not have any children that are also nested.
boolean	isNested() Check if the schema is nested (i.e., top level type is LIST or STRUCT).
boolean	isStructOrHasStructDescendant() Check to see if the schema includes a struct at all.
String	toString()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

INFERRED

public static final Schema INFERRED

Method Detail

getChild

public Schema getChild(int i)

Get the schema of a child element. Note that an inferred schema will have no children.

Parameters:

i - the index of the child to read.

Returns:

the new Schema

Throws:

IndexOutOfBoundsException - if the index is not in the range of children.

toString

public String toString()

Overrides:

toString in class Object

builder

public static Schema.Builder builder()

getFlattenedColumnNames

public String[] getFlattenedColumnNames()

Get names of the columns flattened from all levels in schema by depth-first traversal.

Returns:

An array containing names of all columns in schema.

getColumnNames

public String[] getColumnNames()

Get names of the top level child columns in schema.

Returns:

An array containing names of top level child columns.

isNested

public boolean isNested()

Check if the schema is nested (i.e., top level type is LIST or STRUCT).

Returns:

true if the schema is nested, false otherwise.

hasNestedChildren

public boolean hasNestedChildren()

This is really for a top level struct schema where it is nested, but for things like CSV we care that it does not have any children that are also nested.

getFlattenedTypeIds

public int[] getFlattenedTypeIds()

Get type ids of the columns flattened from all levels in schema by depth-first traversal.

Returns:

An array containing type ids of all columns in schema.

getFlattenedTypeScales

public int[] getFlattenedTypeScales()

Get scales of the columns' types flattened from all levels in schema by depth-first traversal.

Returns:

An array containing type scales of all columns in schema.

getFlattenedDecimalPrecisions

public int[] getFlattenedDecimalPrecisions()

Get decimal precisions of the columns' types flattened from all levels in schema by depth-first traversal.

This is used to pass down the decimal precisions from Spark to only the JNI layer, where some JNI functions require precision values to perform their operations. Decimal precisions should not be consumed by any libcudf's APIs since libcudf does not support precisions for fixed point numbers.

Returns:

An array containing decimal precision of all columns in schema.

getFlattenedTypes

public DType[] getFlattenedTypes()

Get the types of the columns in schema flattened from all levels by depth-first traversal.

Returns:

An array containing types of all columns in schema.

getChildTypes

public DType[] getChildTypes()

Get types of the top level child columns in schema.

Returns:

An array containing types of top level child columns.

getNumChildren

public int getNumChildren()

Get number of top level child columns in schema.

Returns:

Number of child columns.

getFlattenedNumChildren

public int[] getFlattenedNumChildren()

Get numbers of child columns for each level in schema.

Returns:

Numbers of child columns for all levels flattened by depth-first traversal.

getType

public DType getType()

isStructOrHasStructDescendant

public boolean isStructOrHasStructDescendant()

Check to see if the schema includes a struct at all.

Returns:

true if this or any one of its descendants contains a struct, else false.

asHostDataType

public HostColumnVector.DataType asHostDataType()