cuDFOnRayDataframeAxisPartition#

The class is a base class for any axis partition class based on Ray engine and cuDF storage format. This class provides the API to perform operations on an axis partition, using Ray as the execution engine. The axis partition is made up of list of block partitions that are stored in this class.

Public API#

class modin.core.execution.ray.implementations.cudf_on_ray.partitioning.axis_partition.cuDFOnRayDataframeAxisPartition(partitions)#

Base class for any axis partition class for cuDF storage format.

Parameters:: partitions (np.ndarray) – NumPy array with cuDFOnRayDataframePartition-s.

partition_type#: alias of cuDFOnRayDataframePartition

cuOnRayDataframeColumnPartition#

Public API#

class modin.core.execution.ray.implementations.cudf_on_ray.partitioning.axis_partition.cuDFOnRayDataframeColumnPartition(partitions)#

The column partition implementation of cuDFOnRayDataframeAxisPartition.

Parameters:: partitions (np.ndarray) – NumPy array with cuDFOnRayDataframePartition-s.

reduce(func)#

Reduce partitions along self.axis and apply func.

Parameters:: func (callable) – A func to apply.
Return type:: cuDFOnRayDataframePartition

cuDFOnRayDataframeRowPartition#

Public API#

class modin.core.execution.ray.implementations.cudf_on_ray.partitioning.axis_partition.cuDFOnRayDataframeRowPartition(partitions)#

The row partition implementation of cuDFOnRayDataframeAxisPartition.

Parameters:: partitions (np.ndarray) – NumPy array with cuDFOnRayDataframePartition-s.

reduce(func)#

Reduce partitions along self.axis and apply func.

Parameters:: func (calalble) – A func to apply.
Return type:: cuDFOnRayDataframePartition

Notes

Since we are using row partitions, we can bypass the Ray plasma store during axis reduce functions.