databricks/koalas

Popping item from categorical series returns index instead of value

rwedge opened this issue · 3 comments

rwedge commented

This was using koalas 1.8.1

>>> import databricks.koalas as ks
>>> series = ks.Series(['a', 'b', 'c', 'a'])
>>> cat_series = series.astype('category')
>>> series.pop(0)
'a'
>>> cat_series.pop(0)
0
>>> cat_series
1    b
2    c
3    a
dtype: category
Categories (3, object): ['a', 'b', 'c']

I would expect pop to return 'a' in this case, which is what a pandas categorical series would do

HyukjinKwon commented

cc @ueshin @xinrong-databricks FYI

xinrong-meng commented

Thanks for letting us know!

https://issues.apache.org/jira/browse/SPARK-36771 is created for this issue.

  • 👍1