[
https://issues.apache.org/jira/browse/SPARK-59271?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Haotian Sun updated SPARK-59271:
--------------------------------
Component/s: Tests
Affects Version/s: 4.4.0
(was: 5.0.0)
> Add tests for pandas.api.types is_* predicates
> ----------------------------------------------
>
> Key: SPARK-59271
> URL: https://issues.apache.org/jira/browse/SPARK-59271
> Project: Spark
> Issue Type: Sub-task
> Components: PySpark, Tests
> Affects Versions: 4.4.0
> Reporter: Haotian Sun
> Priority: Major
>
> Add golden-file tests pinning the behavior of pandas' type-introspection
> predicates
> (pandas.api.types.is_*) that PySpark depends on.
> PySpark branches on the boolean these predicates return, so an upstream
> change misroutes
> control flow rather than corrupting a single value: the conversion layer
> picks a Spark type
> from is_datetime64_dtype / is_timedelta64_dtype, is_integer_dtype decides
> whether integer
> data is rewritten as Decimal, and pandas-on-Spark dispatches implementations
> and raises
> user-visible TypeErrors from is_hashable / is_list_like.
> The tests run without a Spark session and act as a drift canary, following
> the pattern of
> the existing python/pyspark/tests/upstream/pyarrow suite.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
