[
https://issues.apache.org/jira/browse/SPARK-61936?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated SPARK-61936:
-----------------------------------
Labels: pull-request-available (was: )
> Fix import pyspark.pandas with pandas 3.1 by porting infer_dtype_from_object
> -----------------------------------------------------------------------------
>
> Key: SPARK-61936
> URL: https://issues.apache.org/jira/browse/SPARK-61936
> Project: Spark
> Issue Type: Sub-task
> Components: Pandas API on Spark
> Affects Versions: 5.0.0, 4.4.0
> Reporter: Gabriel Henrique
> Priority: Major
> Labels: pull-request-available
>
> pandas 3.1.0 removed the private helper
> pandas.core.dtypes.common.infer_dtype_from_object, which
> python/pyspark/pandas/frame.py imports at module level. As a result, import
> pyspark.pandas raises ImportError with pandas 3.1.
> Port the helper to pyspark.pandas.utils using only public pandas and NumPy
> APIs, and make DataFrame.select_dtypes use it. Behavior is unchanged: the
> port returns the same result as the pandas helper for every input checked on
> pandas 2.2.3 and 3.0.6.
> This is the remaining half of the pandas 3.1 import failure; the DateOffset
> half was fixed by SPARK-61360.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]