[BEAM-11155] Series.str.repeat zipping operation produces incorrect proxy

• Type: Bug
• Status: Open
• Priority: P3
• Resolution: Unresolved
• Component/s: sdk-py-core
• Assignee: Unassigned
• Reporter: Brian Hulette
• Created: 29/Oct/20 21:41
• Updated: 04/Jun/22 18:14

Description

https://github.com/apache/beam/pull/13139#discussion_r513684704

This proxy is incorrectly inferred as bool.

In [10]: proxy.dtypes
Out[10]: 
str        object
repeats     int64
dtype: object

In [11]: proxy.str.str.repeat(proxy.repeats)
Out[11]: Series([], Name: str, dtype: bool)

The actual operation does produce object though:

In [13]: df.str.str.repeat(df.repeats)
Out[13]: 
0      AAA
1        B
2     CCCC
3    DDDDD
4       EE
Name: str, dtype: object

Currently we work around this by specifying the proxy manually, maybe it can be fixed upstream?