sadpandajoe commented on code in PR #45823:
URL: https://github.com/apache/superset/pull/45823#discussion_r4408160493
##########
pyproject.toml:
##########
@@ -84,11 +84,13 @@ dependencies = [
"marshmallow-union>=0.1.15.post1",
"msgpack>=1.2.3, <1.3",
"nh3>=0.3.7, <0.4",
- "numpy>=1.23.5, <2.5",
+ "numpy>=1.23.5, <2.5; python_version < '3.15'",
+ "numpy>=2.5.4, <2.6; python_version >= '3.15'",
"packaging",
# --------------------------
# pandas and related (wanting pandas[performance] without numba as it's
100+MB and not needed)
- "pandas[excel]>=2.3.3, <2.4",
+ "pandas[excel]>=2.3.3, <2.4; python_version < '3.15'",
+ "pandas[excel]>=3.1.0, <3.2; python_version >= '3.15'",
Review Comment:
Moving 3.15 to pandas 3 makes string columns arrive as the `str` dtype, but
`df_columns_to_num` in `superset/common/utils/dataframe_utils.py` only converts
columns whose `dtype.type == np.object_`. A driver that returns metrics as
strings (the ClickHouse `SUM()` case that helper exists for) now keeps `['10',
'20']` as strings instead of `[10.0, 20.0]`, and the next numeric step (e.g.
the timeseries pivot `mean`) raises `TypeError`. The new boxplot code already
handles this with `is_string_dtype`; should `df_columns_to_num` accept string
dtypes too, with a test that builds the frame with `str` columns?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
