org.bytedeco.arrow

Class VectorKernel

java.lang.Object

org.bytedeco.javacpp.Pointer

org.bytedeco.arrow.Kernel

org.bytedeco.arrow.ArrayKernel

org.bytedeco.arrow.VectorKernel

All Implemented Interfaces:

AutoCloseable

@Namespace(value="arrow::compute")
 @NoOffset
 @Properties(inherit=arrow.class)
public class VectorKernel
extends ArrayKernel

\brief Kernel data structure for implementations of VectorFunction. In addition to the members found in ArrayKernel, contains an optional finalizer function, the null handling and memory pre-allocation preferences (which have different defaults from ScalarKernel), and some other execution-related options.

Nested Class Summary

Nested classes/interfaces inherited from class org.bytedeco.javacpp.Pointer

Pointer.CustomDeallocator, Pointer.Deallocator, Pointer.NativeDeallocator, Pointer.ReferenceCounter

Field Summary

Fields inherited from class org.bytedeco.javacpp.Pointer

address, capacity, limit, position

Constructor Summary

Constructors

Constructor and Description
VectorKernel()
VectorKernel(InputType in_types, OutputType out_type, Pointer exec)
VectorKernel(InputType in_types, OutputType out_type, Pointer exec, Pointer init, Pointer _finalize)
VectorKernel(KernelSignature sig, Pointer exec)
VectorKernel(KernelSignature sig, Pointer exec, Pointer init)
VectorKernel(KernelSignature sig, Pointer exec, Pointer init, Pointer _finalize)
VectorKernel(long size) Native array allocator.
VectorKernel(Pointer p) Pointer cast constructor.

Method Summary

Modifier and Type	Method and Description
Pointer	_finalize() \brief For VectorKernel, convert intermediate results into finalized results.
VectorKernel	_finalize(Pointer setter)
boolean	can_execute_chunkwise() Some vector kernels can do chunkwise execution using ExecBatchIterator, in some cases accumulating some state.
VectorKernel	can_execute_chunkwise(boolean setter)
VectorKernel	getPointer(long i)
MemAllocation.type	mem_allocation()
VectorKernel	mem_allocation(MemAllocation.type setter)
NullHandling.type	null_handling() Since vector kernels generally are implemented rather differently from scalar/elementwise kernels (and they may not even yield arrays of the same size), so we make the developer opt-in to any memory preallocation rather than having to turn it off.
VectorKernel	null_handling(NullHandling.type setter)
boolean	output_chunked() Some kernels (like unique and value_counts) yield non-chunked output from chunked-array inputs.
VectorKernel	output_chunked(boolean setter)
VectorKernel	position(long position)

Methods inherited from class org.bytedeco.arrow.ArrayKernel

can_write_into_slices, can_write_into_slices, exec, exec

Methods inherited from class org.bytedeco.arrow.Kernel

init, init, parallelizable, parallelizable, signature, signature, simd_level, simd_level

Methods inherited from class org.bytedeco.javacpp.Pointer

address, asBuffer, asByteBuffer, availablePhysicalBytes, calloc, capacity, capacity, close, deallocate, deallocate, deallocateReferences, deallocator, deallocator, equals, fill, formatBytes, free, getPointer, getPointer, getPointer, hashCode, isNull, isNull, limit, limit, malloc, maxBytes, maxPhysicalBytes, memchr, memcmp, memcpy, memmove, memset, offsetof, offsetof, parseBytes, physicalBytes, position, put, realloc, referenceCount, releaseReference, retainReference, setNull, sizeof, sizeof, toString, totalBytes, totalCount, totalPhysicalBytes, withDeallocator, zero

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Constructor Detail

VectorKernel

public VectorKernel(Pointer p)

Pointer cast constructor. Invokes Pointer(Pointer).

VectorKernel

public VectorKernel(long size)

Native array allocator. Access with Pointer.position(long).

VectorKernel

public VectorKernel()

VectorKernel

public VectorKernel(@SharedPtr @Cast(value={"","std::shared_ptr<arrow::compute::KernelSignature>"})
                    KernelSignature sig,
                    @ByVal @Cast(value="arrow::compute::ArrayKernelExec*")
                    Pointer exec)

VectorKernel

public VectorKernel(@StdVector
                    InputType in_types,
                    @ByVal
                    OutputType out_type,
                    @ByVal @Cast(value="arrow::compute::ArrayKernelExec*")
                    Pointer exec,
                    @ByVal(nullValue="arrow::compute::KernelInit(nullptr)") @Cast(value="arrow::compute::KernelInit*")
                    Pointer init,
                    @ByVal(nullValue="arrow::compute::VectorFinalize(nullptr)") @Cast(value="arrow::compute::VectorFinalize*")
                    Pointer _finalize)

VectorKernel

public VectorKernel(@StdVector
                    InputType in_types,
                    @ByVal
                    OutputType out_type,
                    @ByVal @Cast(value="arrow::compute::ArrayKernelExec*")
                    Pointer exec)

VectorKernel

public VectorKernel(@SharedPtr @Cast(value={"","std::shared_ptr<arrow::compute::KernelSignature>"})
                    KernelSignature sig,
                    @ByVal @Cast(value="arrow::compute::ArrayKernelExec*")
                    Pointer exec,
                    @ByVal @Cast(value="arrow::compute::KernelInit*")
                    Pointer init,
                    @ByVal(nullValue="arrow::compute::VectorFinalize(nullptr)") @Cast(value="arrow::compute::VectorFinalize*")
                    Pointer _finalize)

VectorKernel

public VectorKernel(@SharedPtr @Cast(value={"","std::shared_ptr<arrow::compute::KernelSignature>"})
                    KernelSignature sig,
                    @ByVal @Cast(value="arrow::compute::ArrayKernelExec*")
                    Pointer exec,
                    @ByVal @Cast(value="arrow::compute::KernelInit*")
                    Pointer init)

Method Detail

position

public VectorKernel position(long position)

Overrides:

position in class ArrayKernel

getPointer

public VectorKernel getPointer(long i)

Overrides:

getPointer in class ArrayKernel

_finalize

@ByRef
 @Cast(value="arrow::compute::VectorFinalize*")
 @Name(value="finalize")
public Pointer _finalize()

\brief For VectorKernel, convert intermediate results into finalized results. Mutates input argument. Some kernels may accumulate state (example: hashing-related functions) through processing chunked inputs, and then need to attach some accumulated state to each of the outputs of processing each chunk of data.

_finalize

public VectorKernel _finalize(Pointer setter)

null_handling

public NullHandling.type null_handling()

Since vector kernels generally are implemented rather differently from scalar/elementwise kernels (and they may not even yield arrays of the same size), so we make the developer opt-in to any memory preallocation rather than having to turn it off.

null_handling

public VectorKernel null_handling(NullHandling.type setter)

mem_allocation

public MemAllocation.type mem_allocation()

mem_allocation

public VectorKernel mem_allocation(MemAllocation.type setter)

can_execute_chunkwise

@Cast(value="bool")
public boolean can_execute_chunkwise()

Some vector kernels can do chunkwise execution using ExecBatchIterator, in some cases accumulating some state. Other kernels (like Take) need to be passed whole arrays and don't work on ChunkedArray inputs

can_execute_chunkwise

public VectorKernel can_execute_chunkwise(boolean setter)

output_chunked

@Cast(value="bool")
public boolean output_chunked()

Some kernels (like unique and value_counts) yield non-chunked output from chunked-array inputs. This option controls how the results are boxed when returned from ExecVectorFunction true -> ChunkedArray false -> Array

output_chunked

public VectorKernel output_chunked(boolean setter)