CSV

class pylibcudf.io.csv.CsvReaderOptions

The settings to use for read_csv For details, see cudf::io::csv_reader_options

Methods

builder(SourceInfo source)	Create a CsvWriterOptionsBuilder object
set_comment(self, str comment)	Sets comment line start character.
set_delimiter(self, str delimiter)	Sets field delimiter.
set_dtypes(self, types)	Sets per-column types.
set_false_values(self, list false_values)	Sets additional values to recognize as boolean false values.
set_header(self, size_type header)	Sets header row index.
set_na_values(self, list na_values)	Sets additional values to recognize as null values.
set_names(self, list col_names)	Sets names of the column.
set_parse_dates(self, list val)	Sets indexes or names of columns to read as datetime.
set_parse_hex(self, list val)	Sets indexes or names of columns to parse as hexadecimal.
set_prefix(self, str prefix)	Sets prefix to be used for column ID.
set_source(self, SourceInfo src)	Set a new source info location.
set_thousands(self, str thousands)	Sets numeric data thousands separator.
set_true_values(self, list true_values)	Sets additional values to recognize as boolean true values.
set_use_cols_indexes(self, list col_indices)	Sets indexes of columns to read.
set_use_cols_names(self, list col_names)	Sets names of the columns to be read.

static builder(SourceInfo source) → CsvReaderOptionsBuilder

Create a CsvWriterOptionsBuilder object

For details, see cudf::io::csv_reader_options::builder()

Parameters:

sinkSourceInfo

The source to read the CSV file from.

Returns:

CsvReaderOptionsBuilder

Builder to build CsvReaderOptions

set_comment(self, str comment) → void

Sets comment line start character.

Parameters:

commentstr

A character that indicates comment

Returns:

None

set_delimiter(self, str delimiter) → void

Sets field delimiter.

Parameters:

delimiterstr

A character to indicate delimiter

Returns:

None

set_dtypes(

self,

types: dict[str, DataType] | list[DataType],

) → void

Sets per-column types.

Parameters:

typesdict[str, data_type] | list[data_type]

Column name to data type map specifying the columns’ target data types. Or a list specifying the columns’ target data types.

Returns:

None

set_false_values(self, list false_values: list[str]) → void

Sets additional values to recognize as boolean false values.

Parameters:

false_valueslist[str]

List of values to be considered to be false

Returns:

None

set_header(self, size_type header) → void

Sets header row index.

Parameters:

headersize_type

Index where header row is located

Returns:

None

set_na_values(self, list na_values: list[str]) → void

Sets additional values to recognize as null values.

Parameters:

na_valueslist[str]

List of values to be considered to be null

Returns:

None

set_names(self, list col_names: list[str]) → void

Sets names of the column.

Parameters:

col_nameslist[str]

List of column names

Returns:

None

set_parse_dates(self, list val: list[int | str]) → void

Sets indexes or names of columns to read as datetime.

Parameters:

vallist[int | str]

List column indices or names to infer as datetime.

Returns:

None

set_parse_hex(self, list val: list[int | str]) → void

Sets indexes or names of columns to parse as hexadecimal.

Parameters:

vallist[int | str]

List of column indices or names to parse as hexadecimal

Returns:

None

set_prefix(self, str prefix) → void

Sets prefix to be used for column ID.

Parameters:

prefixstr

String used as prefix in for each column name

Returns:

None

set_source(self, SourceInfo src) → void

Set a new source info location.

Parameters:

srcSourceInfo

New source information, replacing existing information.

Returns:

None

set_thousands(self, str thousands) → void

Sets numeric data thousands separator.

Parameters:

thousandsstr

A character that separates thousands

Returns:

None

set_true_values(self, list true_values: list[str]) → void

Sets additional values to recognize as boolean true values.

Parameters:

true_valueslist[str]

List of values to be considered to be true

Returns:

None

set_use_cols_indexes(self, list col_indices: list[int]) → void

Sets indexes of columns to read.

Parameters:

col_indiceslist[int]

List of column indices that are needed

Returns:

None

set_use_cols_names(self, list col_names: list[str]) → void

Sets names of the columns to be read.

Parameters:

col_nameslist[str]

List of column indices that are needed

Returns:

None

class pylibcudf.io.csv.CsvReaderOptionsBuilder

Builder to build options for read_csv

For details, see cudf::io::csv_reader_options_builder

Methods

build(self)	Create a CsvReaderOptions object
byte_range_offset(self, size_t byte_range_offset)	Sets number of bytes to skip from source start.
byte_range_size(self, size_t byte_range_size)	Sets number of bytes to read.
compression(self, compression_type compression)	Sets compression format of the source.
dayfirst(self, bool dayfirst)	Sets whether to parse dates as DD/MM versus MM/DD.
decimal(self, str decimal)	Sets decimal point character.
delim_whitespace(self, bool delim_whitespace)	Sets whether to treat whitespace as field delimiter.
delimiter(self, str delimiter)	Sets field delimiter.
doublequote(self, bool doublequote)	Sets a quote inside a value is double-quoted.
keep_default_na(self, bool keep_default_na)	Sets whether to keep the built-in default NA values.
lineterminator(self, str lineterminator)	Sets line terminator.
mangle_dupe_cols(self, bool mangle_dupe_cols)	Sets whether to rename duplicate column names.
na_filter(self, bool na_filter)	Sets whether to disable null filter.
nrows(self, size_type nrows)	Sets number of rows to read.
quotechar(self, str quotechar)	Sets quoting character.
quoting(self, quote_style quoting)	Sets quoting style.
skip_blank_lines(self, bool skip_blank_lines)	Sets whether to ignore empty lines or parse line values as invalid.
skipfooter(self, size_type skipfooter)	Sets number of rows to skip from end.
skipinitialspace(self, bool skipinitialspace)	Sets whether to skip whitespace after the delimiter.
skiprows(self, size_type skiprows)	Sets number of rows to skip from start.

build(self) → CsvReaderOptions

Create a CsvReaderOptions object

byte_range_offset(

self,

size_t byte_range_offset,

) → CsvReaderOptionsBuilder

Sets number of bytes to skip from source start.

Parameters:

byte_range_offsetsize_t

Number of bytes of offset

Returns:

CsvReaderOptionsBuilder

byte_range_size(

self,

size_t byte_range_size,

) → CsvReaderOptionsBuilder

Sets number of bytes to read.

Parameters:

byte_range_offsetsize_t

Number of bytes to read

Returns:

CsvReaderOptionsBuilder

compression(

self,

compression_type compression,

) → CsvReaderOptionsBuilder

Sets compression format of the source.

Parameters:

compressioncompression_type

Compression type

Returns:

CsvReaderOptionsBuilder

dayfirst(

self,

bool dayfirst,

) → CsvReaderOptionsBuilder

Sets whether to parse dates as DD/MM versus MM/DD.

Parameters:

dayfirstbool

Boolean value to enable/disable

Returns:

CsvReaderOptionsBuilder

decimal(

self,

str decimal,

) → CsvReaderOptionsBuilder

Sets decimal point character.

Parameters:

quotecharstr

A character that indicates decimal values

Returns:

CsvReaderOptionsBuilder

delim_whitespace(

self,

bool delim_whitespace,

) → CsvReaderOptionsBuilder

Sets whether to treat whitespace as field delimiter.

Parameters:

delim_whitespacebool

Boolean value to enable/disable

Returns:

CsvReaderOptionsBuilder

delimiter(

self,

str delimiter,

) → CsvReaderOptionsBuilder

Sets field delimiter.

Parameters:

delimiterstr

A character to indicate delimiter

Returns:

CsvReaderOptionsBuilder

doublequote(

self,

bool doublequote,

) → CsvReaderOptionsBuilder

Sets a quote inside a value is double-quoted.

Parameters:

doublequotebool

Boolean value to enable/disable

Returns:

CsvReaderOptionsBuilder

keep_default_na(

self,

bool keep_default_na,

) → CsvReaderOptionsBuilder

Sets whether to keep the built-in default NA values.

Parameters:

keep_default_nabool

Boolean value to enable/disable

Returns:

CsvReaderOptionsBuilder

lineterminator(

self,

str lineterminator,

) → CsvReaderOptionsBuilder

Sets line terminator.

Parameters:

quotingstr

A character to indicate line termination

Returns:

CsvReaderOptionsBuilder

mangle_dupe_cols(

self,

bool mangle_dupe_cols,

) → CsvReaderOptionsBuilder

Sets whether to rename duplicate column names.

Parameters:

mangle_dupe_colsbool

Boolean value to enable/disable

Returns:

CsvReaderOptionsBuilder

na_filter(

self,

bool na_filter,

) → CsvReaderOptionsBuilder

Sets whether to disable null filter.

Parameters:

na_filterbool

Boolean value to enable/disable

Returns:

CsvReaderOptionsBuilder

nrows(

self,

size_type nrows,

) → CsvReaderOptionsBuilder

Sets number of rows to read.

Parameters:

nrowssize_type

Number of rows to read

Returns:

CsvReaderOptionsBuilder

quotechar(

self,

str quotechar,

) → CsvReaderOptionsBuilder

Sets quoting character.

Parameters:

quotecharstr

A character to indicate quoting

Returns:

CsvReaderOptionsBuilder

quoting(

self,

quote_style quoting,

) → CsvReaderOptionsBuilder

Sets quoting style.

Parameters:

quotingquote_style

Quoting style used

Returns:

CsvReaderOptionsBuilder

skip_blank_lines(

self,

bool skip_blank_lines,

) → CsvReaderOptionsBuilder

Sets whether to ignore empty lines or parse line values as invalid.

Parameters:

skip_blank_linesbool

Boolean value to enable/disable

Returns:

CsvReaderOptionsBuilder

skipfooter(

self,

size_type skipfooter,

) → CsvReaderOptionsBuilder

Sets number of rows to skip from end.

Parameters:

skipfootersize_type

Number of rows to skip

Returns:

CsvReaderOptionsBuilder

skipinitialspace(

self,

bool skipinitialspace,

) → CsvReaderOptionsBuilder

Sets whether to skip whitespace after the delimiter.

Parameters:

skipinitialspacebool

Boolean value to enable/disable

Returns:

CsvReaderOptionsBuilder

skiprows(

self,

size_type skiprows,

) → CsvReaderOptionsBuilder

Sets number of rows to skip from start.

Parameters:

skiprowssize_type

Number of rows to skip

Returns:

CsvReaderOptionsBuilder

class pylibcudf.io.csv.CsvWriterOptions

The settings to use for write_csv

For details, see cudf::io::csv_writer_options

Methods

builder(SinkInfo sink, Table table)

Create a CsvWriterOptionsBuilder object

static builder(

SinkInfo sink,

Table table,

) → CsvWriterOptionsBuilder

Create a CsvWriterOptionsBuilder object

For details, see cudf::io::csv_writer_options::builder()

Parameters:

sinkSinkInfo

The sink used for writer output

tableTable

Table to be written to output

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

class pylibcudf.io.csv.CsvWriterOptionsBuilder

Builder to build options for write_csv

For details, see cudf::io::csv_writer_options_builder

Methods

build(self)	Create a CsvWriterOptions object
compression(self, compression_type comp)	Sets the compression type for the output
compression_block_size(self, size_t size)	Sets the size of the blocks that the output is compressed in
false_value(self, str val)	Sets string used for values == 0
include_header(self, bool val)	Enables/Disables headers being written to csv.
inter_column_delimiter(self, str delim)	Sets character used for separating column values.
line_terminator(self, str term)	Sets character used for separating lines.
na_rep(self, str val)	Sets string to used for null entries.
names(self, list names)	Sets optional column names.
quoting(self, quote_style style)	Sets quoting style for string columns.
rows_per_chunk(self, int val)	Sets maximum number of rows to process for each file write.
true_value(self, str val)	Sets string used for values != 0

build(self) → CsvWriterOptions

Create a CsvWriterOptions object

compression(

self,

compression_type comp,

) → CsvWriterOptionsBuilder

Sets the compression type for the output

Parameters:

compCompressionType

Compression type. Only NONE and ZSTD are supported for CSV writer. ZSTD is used because it supports concatenated frames, enabling progressive chunk-based compression compatible with standard decompression tools.

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

compression_block_size(

self,

size_t size,

) → CsvWriterOptionsBuilder

Sets the size of the blocks that the output is compressed in

The output is split into blocks of this size, each compressed into its own frame, so that the codec can compress them in parallel. The block size is independent of rows_per_chunk, and is capped at the maximum input size the codec supports.

Parameters:

sizeint

Compression block size, in bytes. Must be greater than zero.

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

false_value(

self,

str val,

) → CsvWriterOptionsBuilder

Sets string used for values == 0

Parameters:

valstr

String to represent values == 0

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

include_header(

self,

bool val,

) → CsvWriterOptionsBuilder

Enables/Disables headers being written to csv.

Parameters:

valbool

Boolean value to enable/disable

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

inter_column_delimiter(

self,

str delim,

) → CsvWriterOptionsBuilder

Sets character used for separating column values.

Parameters:

delimstr

Character to delimit column values

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

line_terminator(

self,

str term,

) → CsvWriterOptionsBuilder

Sets character used for separating lines.

Parameters:

termstr

Character to represent line termination

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

na_rep(

self,

str val,

) → CsvWriterOptionsBuilder

Sets string to used for null entries.

Parameters:

valstr

String to represent null value

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

names(

self,

list names,

) → CsvWriterOptionsBuilder

Sets optional column names.

Parameters:

nameslist[str]

Column names

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

quoting(

self,

quote_style style,

) → CsvWriterOptionsBuilder

Sets quoting style for string columns.

Parameters:

stylequote_style

Quoting style (MINIMAL or NONE). MINIMAL: Quote only fields containing special characters. NONE: Never quote fields.

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

rows_per_chunk(

self,

int val,

) → CsvWriterOptionsBuilder

Sets maximum number of rows to process for each file write.

Parameters:

valint

Number of rows per chunk

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

true_value(

self,

str val,

) → CsvWriterOptionsBuilder

Sets string used for values != 0

Parameters:

valstr

String to represent values != 0

Returns:

CsvWriterOptionsBuilder

Builder to build CsvWriterOptions

pylibcudf.io.csv.read_csv(

CsvReaderOptions options,

stream: CudaStreamLike | None = None,

DeviceMemoryResource mr=None,

) → TableWithMetadata

Read from CSV format.

The source to read from and options are encapsulated by the options object.

For details, see read_csv().

Parameters:

options: CsvReaderOptions

Settings for controlling reading behavior

streamStream | None

CUDA stream used for device memory operations and kernel launches

mrDeviceMemoryResource, optional

Device memory resource used to allocate the returned table’s device memory.

pylibcudf.io.csv.write_csv(

CsvWriterOptions options,

stream: CudaStreamLike | None = None,

) → void

Write to CSV format.

The table to write, output paths, and options are encapsulated by the options object.

For details, see write_csv().

Parameters:

options: CsvWriterOptions

Settings for controlling writing behavior

streamStream | None

CUDA stream used for device memory operations and kernel launches