[ARROW-18137] [Python][Docs] Allow passing no aggregations to TableGroupBy.aggregate

• Type: New Feature
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 9.0.0
• Fix Version/s: 11.0.0
• Component/s: Documentation, Python
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/33333
• Assignee: Jacek Pliszka
• Reporter: Jacek Pliszka
• Created: 23/Oct/22 18:26
• Updated: 11/Jan/23 11:58
• Resolved: 24/Oct/22 21:53

Description

If we could allow TableGroupBy.aggregate to accept no aggregation functions then it would behave like pandas drop_duplicates.

t.group_by(['keys', 'values']).aggregate()

I did some naive benchmarks and looks like it should be 30% faster than converting to pandas and deduplicating. This was my naive test:

 t.append_column('i', pa.array([1]*len(t),pa.int64())).group_by(['keys', 'values']).aggregate([("i", "max")]).drop(['i_max'])

And on small 5M table it took 245ms while 359ms for t.to_pandas().drop_duplicates()

Actual aggregation without adding dummy column should be  even faster still will allow drop_duplicates functionality until better implementation arrives

Issue Links

• links to: GitHub Pull Request #14482