[ARROW-7709] [Python] Conversion from Table Column to Pandas loses name for Timestamps

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.16.0
• Component/s: Python
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/23952
• Assignee: Joris Van den Bossche
• Reporter: Bryan Cutler
• Created: 28/Jan/20 22:25
• Updated: 11/Jan/23 07:55
• Resolved: 30/Jan/20 10:41

Description

When converting a Table timestamp column to Pandas, the name of the column is lost in the resulting series.

In [23]: a1 = pa.array([pd.Timestamp.now()])                                                                                 

In [24]: a2 = pa.array([1])                                                                                                  

In [25]: t = pa.Table.from_arrays([a1, a2], ['ts', 'a'])                                                                     

In [26]: for c in t: 
    ...:     print(c.to_pandas()) 
    ...:                                                                                                                     
0   2020-01-28 13:17:26.738708
dtype: datetime64[ns]
0    1
Name: a, dtype: int64 

Issue Links

• is related to: ARROW-7693 [CI] Fix test-conda-python-3.7-spark-master nightly errors
• links to: GitHub Pull Request #6312