null_mask#

pylibcudf.null_mask.bitmask_allocation_size_bytes(size_type number_of_bits) → size_t#

Computes the required bytes necessary to represent the specified number of bits with a 64B padding boundary.

For details, see bitmask_allocation_size_bytes().

Parameters:

number_of_bitssize_type: The number of bits that need to be represented

Returns:

size_t: The necessary number of bytes

pylibcudf.null_mask.bitmask_and(list columns, Stream stream=None, DeviceMemoryResource mr=None) → tuple#

Performs bitwise AND of the bitmasks of a list of columns.

For details, see bitmask_and().

Parameters:

columnslist: The list of columns
streamStream | None: CUDA stream on which to perform the operation.
mrDeviceMemoryResource | None: Device memory resource for allocations.

Returns:

tuple[DeviceBuffer, size_type]: A tuple of the resulting mask and count of unset bits

pylibcudf.null_mask.bitmask_or(list columns, Stream stream=None, DeviceMemoryResource mr=None) → tuple#

Performs bitwise OR of the bitmasks of a list of columns.

For details, see bitmask_or().

Parameters:

columnslist: The list of columns
streamStream | None: CUDA stream on which to perform the operation.
mrDeviceMemoryResource | None: Device memory resource for allocations.

Returns:

tuple[DeviceBuffer, size_type]: A tuple of the resulting mask and count of unset bits

pylibcudf.null_mask.copy_bitmask(Column col, Stream stream=None, DeviceMemoryResource mr=None) → rmm.DeviceBuffer#

Copies col’s bitmask into a DeviceBuffer.

For details, see copy_bitmask().

Parameters:

colColumn: Column whose bitmask needs to be copied
streamStream | None: CUDA stream on which to perform the operation.
mrDeviceMemoryResource | None: Device memory resource for allocations.

Returns:

rmm.DeviceBuffer: A DeviceBuffer containing col’s bitmask, or an empty DeviceBuffer if col is not nullable

pylibcudf.null_mask.create_null_mask(size_type size, mask_state state=mask_state.UNINITIALIZED, Stream stream=None, DeviceMemoryResource mr=None) → rmm.DeviceBuffer#

Creates a DeviceBuffer for use as a null value indicator bitmask of a Column.

For details, see create_null_mask().

Parameters:

sizesize_type: The number of elements to be represented by the mask
statemask_state, optional: The desired state of the mask. Can be one of { MaskState.UNALLOCATED, MaskState.UNINITIALIZED, MaskState.ALL_VALID, MaskState.ALL_NULL } (default MaskState.UNINITIALIZED)
streamStream | None: CUDA stream on which to perform the operation.
mrDeviceMemoryResource | None: Device memory resource for allocations.

Returns:

rmm.DeviceBuffer: A DeviceBuffer for use as a null bitmask satisfying the desired size and state

pylibcudf.null_mask.null_count(bitmask, size_type start, size_type stop, Stream stream=None) → size_type#

Given a validity bitmask, counts the number of null elements.

For details, see null_count().

Parameters:

bitmaskSpan-like object: Object with ptr and size attributes (e.g., gpumemoryview, Buffer, DeviceBuffer).
startint: Index of the first bit to count (inclusive).
stopint: Index of the last bit to count (exclusive).
streamStream | None: CUDA stream on which to perform the operation.

Returns:

int: The number of null elements in the specified range.

null_mask#

This Page