pyarrow.parquet.write_to_dataset

pyarrow.parquet.write_to_dataset(table, root_path, partition_cols=None, partition_filename_cb=None, filesystem=None, use_legacy_dataset=None, schema=None, partitioning=None, basename_template=None, use_threads=None, file_visitor=None, existing_data_behavior=None, **kwargs)[source]

Wrapper around dataset.write_dataset (when use_legacy_dataset=False) or parquet.write_table (when use_legacy_dataset=True) for writing a Table to Parquet format by partitions. For each combination of partition columns and values, a subdirectories are created in the following manner:

root_dir/
group1=value1
group2=value1

<uuid>.parquet

group2=value2

<uuid>.parquet

group1=valueN
group2=value1

<uuid>.parquet

group2=valueN

<uuid>.parquet

Parameters:
tablepyarrow.Table
root_pathstr, pathlib.Path

The root directory of the dataset

filesystemFileSystem, default None

If nothing passed, will be inferred based on path. Path will try to be found in the local on-disk filesystem otherwise it will be parsed as an URI to determine the filesystem.

partition_colslist,

Column names by which to partition the dataset. Columns are partitioned in the order they are given

partition_filename_cbcallable,

A callback function that takes the partition key(s) as an argument and allow you to override the partition filename. If nothing is passed, the filename will consist of a uuid. This option is only supported for use_legacy_dataset=True. When use_legacy_dataset=None and this option is specified, use_legacy_datase will be set to True.

use_legacy_datasetbool

Default is False. Set to True to use the the legacy behaviour (this option is deprecated, and the legacy implementation will be removed in a future version). The legacy implementation still supports the partition_filename_cb keyword but is less efficient when using partition columns.

use_threadsbool, default True

Write files in parallel. If enabled, then maximum parallelism will be used determined by the number of available CPU cores. This option is only supported for use_legacy_dataset=False.

schemaSchema, optional

This option is only supported for use_legacy_dataset=False.

partitioningPartitioning or list[str], optional

The partitioning scheme specified with the pyarrow.dataset.partitioning() function or a list of field names. When providing a list of field names, you can use partitioning_flavor to drive which partitioning type should be used. This option is only supported for use_legacy_dataset=False.

basename_templatestr, optional

A template string used to generate basenames of written data files. The token ‘{i}’ will be replaced with an automatically incremented integer. If not specified, it defaults to “guid-{i}.parquet”. This option is only supported for use_legacy_dataset=False.

file_visitorfunction

If set, this function will be called with a WrittenFile instance for each file created during the call. This object will have both a path attribute and a metadata attribute.

The path attribute will be a string containing the path to the created file.

The metadata attribute will be the parquet metadata of the file. This metadata will have the file path attribute set and can be used to build a _metadata file. The metadata attribute will be None if the format is not parquet.

Example visitor which simple collects the filenames created:

visited_paths = []

def file_visitor(written_file):
    visited_paths.append(written_file.path)

This option is only supported for use_legacy_dataset=False.

existing_data_behavior‘overwrite_or_ignore’ | ‘error’ | ‘delete_matching’

Controls how the dataset will handle data that already exists in the destination. The default behaviour is ‘overwrite_or_ignore’.

‘overwrite_or_ignore’ will ignore any existing data and will overwrite files with the same name as an output file. Other existing files will be ignored. This behavior, in combination with a unique basename_template for each write, will allow for an append workflow.

‘error’ will raise an error if any data exists in the destination.

‘delete_matching’ is useful when you are writing a partitioned dataset. The first time each partition directory is encountered the entire directory will be deleted. This allows you to overwrite old partitions completely. This option is only supported for use_legacy_dataset=False.

**kwargsdict,

When use_legacy_dataset=False, used as additional kwargs for dataset.write_dataset function (passed to ParquetFileFormat.make_write_options). See the docstring of write_table for the available options. When use_legacy_dataset=True, used as additional kwargs for parquet.write_table function (See docstring for write_table or ParquetWriter for more information). Using metadata_collector in kwargs allows one to collect the file metadata instances of dataset pieces. The file paths in the ColumnChunkMetaData will be set relative to root_path.

Examples

Generate an example PyArrow Table:

>>> import pyarrow as pa
>>> table = pa.table({'year': [2020, 2022, 2021, 2022, 2019, 2021],
...                   'n_legs': [2, 2, 4, 4, 5, 100],
...                   'animal': ["Flamingo", "Parrot", "Dog", "Horse",
...                              "Brittle stars", "Centipede"]})

and write it to a partitioned dataset:

>>> import pyarrow.parquet as pq
>>> pq.write_to_dataset(table, root_path='dataset_name_3',
...                     partition_cols=['year'])
>>> pq.ParquetDataset('dataset_name_3', use_legacy_dataset=False).files
['dataset_name_3/year=2019/...-0.parquet', ...

Write a single Parquet file into the root folder:

>>> pq.write_to_dataset(table, root_path='dataset_name_4')
>>> pq.ParquetDataset('dataset_name_4/', use_legacy_dataset=False).files
['dataset_name_4/...-0.parquet']