pyarrow.cuda.BufferWriter¶

class pyarrow.cuda.BufferWriter¶

Bases: pyarrow.lib.NativeFile

File interface for writing to CUDA buffers.

By default writes are unbuffered. Use set_buffer_size to enable buffering.

__init__(*args, **kwargs)¶

Methods

`__init__`(args, *kwargs)
`close`(self)
`download`(self, stream_or_path[, buffer_size])	Read file completely to local path (rather than reading completely into memory).
`fileno`(self)	NOT IMPLEMENTED
`flush`(self)	Flush the buffer stream
`isatty`(self)
`metadata`(self)	Return file metadata
`read`(self[, nbytes])	Read indicated number of bytes from file, or read all remaining bytes if no argument passed
`read1`(self[, nbytes])	Read and return up to n bytes.
`read_at`(self, nbytes, offset)	Read indicated number of bytes at offset from the file
`read_buffer`(self[, nbytes])
`readable`(self)
`readall`(self)
`readinto`(self, b)	Read into the supplied buffer
`readline`(self[, size])	NOT IMPLEMENTED.
`readlines`(self[, hint])	NOT IMPLEMENTED.
`seek`(self, int64_t position, int whence=0)
`seekable`(self)
`size`(self)	Return file size
`tell`(self)	Return current stream position
`truncate`(self)	NOT IMPLEMENTED
`upload`(self, stream[, buffer_size])	Pipe file-like object to file
`writable`(self)
`write`(self, data)	Write byte from any object implementing buffer protocol (bytes, bytearray, ndarray, pyarrow.Buffer)
`writeat`(self, int64_t position, data)	Write data to buffer starting from position.
`writelines`(self, lines)

Attributes

`buffer_size`	Returns size of host (CPU) buffer, 0 for unbuffered
`closed`
`mode`	The file mode.
`num_bytes_buffered`	Returns number of bytes buffered on host

buffer_size¶: Returns size of host (CPU) buffer, 0 for unbuffered

close(self)¶

closed¶

download(self, stream_or_path, buffer_size=None)¶: Read file completely to local path (rather than reading completely into memory). First seeks to the beginning of the file.

fileno(self)¶: NOT IMPLEMENTED

flush(self)¶: Flush the buffer stream

isatty(self)¶

metadata(self)¶: Return file metadata

mode¶

The file mode. Currently instances of NativeFile may support:

rb: binary read
wb: binary write
rb+: binary read and write

num_bytes_buffered¶: Returns number of bytes buffered on host

read(self, nbytes=None)¶

Read indicated number of bytes from file, or read all remaining bytes if no argument passed

Parameters

nbytesint, default None

Returns

databytes

read1(self, nbytes=None)¶

Read and return up to n bytes.

Alias for read, needed to match the IOBase interface.

read_at(self, nbytes, offset)¶

Read indicated number of bytes at offset from the file

Parameters

nbytesint
offsetint

Returns

databytes

read_buffer(self, nbytes=None)¶

readable(self)¶

readall(self)¶

readinto(self, b)¶

Read into the supplied buffer

Parameters

b: any python object supporting buffer interface

Returns

int: number of bytes written

readline(self, size=None)¶

NOT IMPLEMENTED. Read and return a line of bytes from the file.

If size is specified, read at most size bytes.

Line terminator is always b”n”.

readlines(self, hint=None)¶

NOT IMPLEMENTED. Read lines of the file

Parameters

hint: int maximum number of bytes read until we stop

seek(self, int64_t position, int whence=0)¶

seekable(self)¶

size(self)¶: Return file size

tell(self)¶: Return current stream position

truncate(self)¶: NOT IMPLEMENTED

upload(self, stream, buffer_size=None)¶: Pipe file-like object to file

writable(self)¶

write(self, data)¶

Write byte from any object implementing buffer protocol (bytes, bytearray, ndarray, pyarrow.Buffer)

Parameters

databytes-like object or exporter of buffer protocol

Returns

int: nbytes: number of bytes written

writeat(self, int64_t position, data)¶

Write data to buffer starting from position.

Parameters

positionint: Specify device buffer position where the data will be written.
dataarray-like: Specify data, the data instance must implement buffer protocol.

writelines(self, lines)¶

pyarrow.cuda.BufferReader

pyarrow.cuda.serialize_record_batch