[ARROW-5258] [C++/Python] Expose file metadata of dataset pieces to caller

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.14.0
• Component/s: C++, Python
• Labels: parquet pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/21729
• Assignee: Pearu Peterson
• Reporter: Pearu Peterson
• Created: 04/May/19 18:27
• Updated: 11/Jan/23 07:39
• Resolved: 06/May/19 16:45

Description

This issue resolves partly the issue raised in ARROW-1983 by providing a way to collect FileMetaData objects of dataset pieces created via `write_to_dataset` function:

metadata_list = []
write_to_dataset(table, ..., metadata_collector=metadata_list)

where `metadata_list` will contain instances of file metadata objects of dataset pieces.

Issue Links

• relates to: ARROW-1983 [Python] Add ability to write parquet `_metadata` file
• links to: GitHub Pull Request #4236
• links to: GitHub Pull Request #4236