[ARROW-6568] [C++][Python][Parquet] pyarrow.parquet crash writing zero-chunk dictionary-type column

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.14.1
• Fix Version/s: 0.15.0
• Component/s: Python
• Labels: pull-request-available
• Environment: Pyarrow v0.14.1, manylinux1
• External issue URL: https://github.com/apache/arrow/issues/22926
• Assignee: Ben Kietzman
• Reporter: Adam Hooper
• Created: 15/Sep/19 18:47
• Updated: 11/Jan/23 07:47
• Resolved: 17/Sep/19 17:14

Description

Trying to write a zero-RecordBatch file to parquet:

import pyarrow
import pyarrow.parquet
table = pyarrow.Table.from_batches([], pyarrow.schema([('A', pyarrow.dictionary(pyarrow.int32(), pyarrow.string()))]))
pyarrow.parquet.write_table(table, 'x.parquet')

... I receive an error and Python exits with exit code 139:

WARNING: Logging before InitGoogleLogging() is written to STDERR
F0915 18:37:23.099939     1 table.cc:64]  Check failed: (chunks.size()) > (0) cannot construct ChunkedArray from empty vector and omitted type
*** Check failure stack trace: ***

Issue Links

• links to: GitHub Pull Request #5393