[ARROW-1732] [Python] RecordBatch.from_pandas fails on DataFrame with no columns when preserve_index=False

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.8.0
• Component/s: Python
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/17737
• Assignee: Wes McKinney
• Reporter: Wes McKinney
• Created: 25/Oct/17 19:09
• Updated: 11/Jan/23 07:16
• Resolved: 26/Oct/17 12:16

Description

I believe this should have well-defined behavior and not raise an error:

In [5]: pa.RecordBatch.from_pandas(pd.DataFrame({}), preserve_index=False)
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-5-4dda72b47dbd> in <module>()
----> 1 pa.RecordBatch.from_pandas(pd.DataFrame({}), preserve_index=False)

~/code/arrow/python/pyarrow/table.pxi in pyarrow.lib.RecordBatch.from_pandas (/home/wesm/code/arrow/python/build/temp.linux-x86_64-3.5/lib.cxx:39957)()
    586             df, schema, preserve_index, nthreads=nthreads
    587         )
--> 588         return cls.from_arrays(arrays, names, metadata)
    589 
    590     @staticmethod

~/code/arrow/python/pyarrow/table.pxi in pyarrow.lib.RecordBatch.from_arrays (/home/wesm/code/arrow/python/build/temp.linux-x86_64-3.5/lib.cxx:40130)()
    615 
    616         if not number_of_arrays:
--> 617             raise ValueError('Record batch cannot contain no arrays (for now)')
    618 
    619         num_rows = len(arrays[0])

ValueError: Record batch cannot contain no arrays (for now)

Issue Links

• links to: GitHub Pull Request #1252