pyarrow.csv.ConvertOptions

class pyarrow.csv.ConvertOptions(check_utf8=None, *, column_types=None, null_values=None, true_values=None, false_values=None, decimal_point=None, strings_can_be_null=None, quoted_strings_can_be_null=None, include_columns=None, include_missing_columns=None, auto_dict_encode=None, auto_dict_max_cardinality=None, timestamp_parsers=None)

Bases: _Weakrefable

Options for converting CSV data.

Parameters:

check_utf8bool, optional (default True)

Whether to check UTF8 validity of string columns.

column_typespyarrow.Schema or dict, optional

Explicitly map column names to column types. Passing this argument disables type inference on the defined columns.

null_valueslist, optional

A sequence of strings that denote nulls in the data (defaults are appropriate in most cases). Note that by default, string columns are not checked for null values. To enable null checking for those, specify strings_can_be_null=True.

true_valueslist, optional

A sequence of strings that denote true booleans in the data (defaults are appropriate in most cases).

false_valueslist, optional

A sequence of strings that denote false booleans in the data (defaults are appropriate in most cases).

decimal_point1-character str, optional (default ‘.’)

The character used as decimal point in floating-point and decimal data.

strings_can_be_nullbool, optional (default False)

Whether string / binary columns can have null values. If true, then strings in null_values are considered null for string columns. If false, then all strings are valid string values.

quoted_strings_can_be_nullbool, optional (default True)

Whether quoted values can be null. If true, then strings in “null_values” are also considered null when they appear quoted in the CSV file. Otherwise, quoted values are never considered null.

include_columnslist, optional

The names of columns to include in the Table. If empty, the Table will include all columns from the CSV file. If not empty, only these columns will be included, in this order.

include_missing_columnsbool, optional (default False)

If false, columns in include_columns but not in the CSV file will error out. If true, columns in include_columns but not in the CSV file will produce a column of nulls (whose type is selected using column_types, or null by default). This option is ignored if include_columns is empty.

auto_dict_encodebool, optional (default False)

Whether to try to automatically dict-encode string / binary data. If true, then when type inference detects a string or binary column, it it dict-encoded up to auto_dict_max_cardinality distinct values (per chunk), after which it switches to regular encoding. This setting is ignored for non-inferred columns (those in column_types).

auto_dict_max_cardinalityint, optional

The maximum dictionary cardinality for auto_dict_encode. This value is per chunk.

timestamp_parserslist, optional

A sequence of strptime()-compatible format strings, tried in order when attempting to infer or convert timestamp values (the special value ISO8601() can also be given). By default, a fast built-in ISO-8601 parser is used.

Examples

Defining an example data:

>>> import io
>>> s = (
...     "animals,n_legs,entry,fast\n"
...     "Flamingo,2,01/03/2022,Yes\n"
...     "Horse,4,02/03/2022,Yes\n"
...     "Brittle stars,5,03/03/2022,No\n"
...     "Centipede,100,04/03/2022,No\n"
...     ",6,05/03/2022,"
... )
>>> print(s)
animals,n_legs,entry,fast
Flamingo,2,01/03/2022,Yes
Horse,4,02/03/2022,Yes
Brittle stars,5,03/03/2022,No
Centipede,100,04/03/2022,No
,6,05/03/2022,

Change the type of a column:

>>> import pyarrow as pa
>>> from pyarrow import csv
>>> convert_options = csv.ConvertOptions(column_types={"n_legs": pa.float64()})
>>> csv.read_csv(io.BytesIO(s.encode()), convert_options=convert_options)
pyarrow.Table
animals: string
n_legs: double
entry: string
fast: string
----
animals: [["Flamingo","Horse","Brittle stars","Centipede",""]]
n_legs: [[2,4,5,100,6]]
entry: [["01/03/2022","02/03/2022","03/03/2022","04/03/2022","05/03/2022"]]
fast: [["Yes","Yes","No","No",""]]

Define a date parsing format to get a timestamp type column (in case dates are not in ISO format and not converted by default):

>>> convert_options = csv.ConvertOptions(
...                   timestamp_parsers=["%m/%d/%Y", "%m-%d-%Y"])
>>> csv.read_csv(io.BytesIO(s.encode()), convert_options=convert_options)
pyarrow.Table
animals: string
n_legs: int64
entry: timestamp[s]
fast: string
----
animals: [["Flamingo","Horse","Brittle stars","Centipede",""]]
n_legs: [[2,4,5,100,6]]
entry: [[2022-01-03 00:00:00,2022-02-03 00:00:00,2022-03-03 00:00:00,2022-04-03 00:00:00,2022-05-03 00:00:00]]
fast: [["Yes","Yes","No","No",""]]

Specify a subset of columns to be read:

>>> convert_options = csv.ConvertOptions(
...                   include_columns=["animals", "n_legs"])
>>> csv.read_csv(io.BytesIO(s.encode()), convert_options=convert_options)
pyarrow.Table
animals: string
n_legs: int64
----
animals: [["Flamingo","Horse","Brittle stars","Centipede",""]]
n_legs: [[2,4,5,100,6]]

List additional column to be included as a null typed column:

>>> convert_options = csv.ConvertOptions(
...                   include_columns=["animals", "n_legs", "location"],
...                   include_missing_columns=True)
>>> csv.read_csv(io.BytesIO(s.encode()), convert_options=convert_options)
pyarrow.Table
animals: string
n_legs: int64
location: null
----
animals: [["Flamingo","Horse","Brittle stars","Centipede",""]]
n_legs: [[2,4,5,100,6]]
location: [5 nulls]

Define columns as dictionary type (by default only the string/binary columns are dictionary encoded):

>>> convert_options = csv.ConvertOptions(
...                   timestamp_parsers=["%m/%d/%Y", "%m-%d-%Y"],
...                   auto_dict_encode=True)
>>> csv.read_csv(io.BytesIO(s.encode()), convert_options=convert_options)
pyarrow.Table
animals: dictionary<values=string, indices=int32, ordered=0>
n_legs: int64
entry: timestamp[s]
fast: dictionary<values=string, indices=int32, ordered=0>
----
animals: [  -- dictionary:
["Flamingo","Horse","Brittle stars","Centipede",""]  -- indices:
[0,1,2,3,4]]
n_legs: [[2,4,5,100,6]]
entry: [[2022-01-03 00:00:00,2022-02-03 00:00:00,2022-03-03 00:00:00,2022-04-03 00:00:00,2022-05-03 00:00:00]]
fast: [  -- dictionary:
["Yes","No",""]  -- indices:
[0,0,1,1,2]]

Set upper limit for the number of categories. If the categories is more than the limit, the conversion to dictionary will not happen:

>>> convert_options = csv.ConvertOptions(
...                   include_columns=["animals"],
...                   auto_dict_encode=True,
...                   auto_dict_max_cardinality=2)
>>> csv.read_csv(io.BytesIO(s.encode()), convert_options=convert_options)
pyarrow.Table
animals: string
----
animals: [["Flamingo","Horse","Brittle stars","Centipede",""]]

Set empty strings to missing values:

>>> convert_options = csv.ConvertOptions(include_columns=["animals", "n_legs"],
...                   strings_can_be_null=True)
>>> csv.read_csv(io.BytesIO(s.encode()), convert_options=convert_options)
pyarrow.Table
animals: string
n_legs: int64
----
animals: [["Flamingo","Horse","Brittle stars","Centipede",null]]
n_legs: [[2,4,5,100,6]]

Define values to be True and False when converting a column into a bool type:

>>> convert_options = csv.ConvertOptions(
...                   include_columns=["fast"],
...                   false_values=["No"],
...                   true_values=["Yes"])
>>> csv.read_csv(io.BytesIO(s.encode()), convert_options=convert_options)
pyarrow.Table
fast: bool
----
fast: [[true,true,false,false,null]]

__init__(*args, **kwargs)

Methods

__init__(*args, **kwargs)
equals(self, ConvertOptions other)
validate(self)

Attributes

auto_dict_encode	Whether to try to automatically dict-encode string / binary data.
auto_dict_max_cardinality	The maximum dictionary cardinality for auto_dict_encode.
check_utf8	Whether to check UTF8 validity of string columns.
column_types	Explicitly map column names to column types.
decimal_point	The character used as decimal point in floating-point and decimal data.
false_values	A sequence of strings that denote false booleans in the data.
include_columns	The names of columns to include in the Table.
include_missing_columns	If false, columns in include_columns but not in the CSV file will error out.
null_values	A sequence of strings that denote nulls in the data.
quoted_strings_can_be_null	Whether quoted values can be null.
strings_can_be_null	Whether string / binary columns can have null values.
timestamp_parsers	A sequence of strptime()-compatible format strings, tried in order when attempting to infer or convert timestamp values (the special value ISO8601() can also be given).
true_values	A sequence of strings that denote true booleans in the data.

auto_dict_encode

Whether to try to automatically dict-encode string / binary data.

auto_dict_max_cardinality

The maximum dictionary cardinality for auto_dict_encode.

This value is per chunk.

check_utf8

Whether to check UTF8 validity of string columns.

column_types

Explicitly map column names to column types.

decimal_point

The character used as decimal point in floating-point and decimal data.

equals(self, ConvertOptions other)

Parameters:

otherpyarrow.csv.ConvertOptions

Returns:

bool

false_values

A sequence of strings that denote false booleans in the data.

include_columns

The names of columns to include in the Table.

If empty, the Table will include all columns from the CSV file. If not empty, only these columns will be included, in this order.

include_missing_columns

If false, columns in include_columns but not in the CSV file will error out. If true, columns in include_columns but not in the CSV file will produce a null column (whose type is selected using column_types, or null by default). This option is ignored if include_columns is empty.

null_values

A sequence of strings that denote nulls in the data.

quoted_strings_can_be_null

Whether quoted values can be null.

strings_can_be_null

Whether string / binary columns can have null values.

timestamp_parsers

A sequence of strptime()-compatible format strings, tried in order when attempting to infer or convert timestamp values (the special value ISO8601() can also be given). By default, a fast built-in ISO-8601 parser is used.

true_values

A sequence of strings that denote true booleans in the data.

validate(self)