[
https://issues.apache.org/jira/browse/SPARK-23314?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16349897#comment-16349897
]
Felix Cheung commented on SPARK-23314:
--------------------------------------
code
>>> flights = spark.read.option("inferSchema", True).option("header",
>>> True).option("dateFormat", "yyyy-MM-dd HH:mm:ss").csv("data*.csv")
>>> from pyspark.sql.functions import pandas_udf, PandasUDFType
>>> @pandas_udf(flights.schema, PandasUDFType.GROUPED_MAP)
... def subtract_mean_year_mfr(pdf):
... return pdf.assign(year_mfr=pdf.year_mfr - pdf.year_mfr.mean())
...
g = flights.groupby('mfr').apply(subtract_mean_year_mfr)
>>> g = flights.groupby('mfr').apply(subtract_mean_year_mfr)
>>>
>>> g.count()
> Pandas grouped udf on dataset with timestamp column error
> ----------------------------------------------------------
>
> Key: SPARK-23314
> URL: https://issues.apache.org/jira/browse/SPARK-23314
> Project: Spark
> Issue Type: Bug
> Components: PySpark
> Affects Versions: 2.3.0
> Reporter: Felix Cheung
> Priority: Major
>
> Under SPARK-22216
> When testing pandas_udf on group bys, I saw this error with the timestamp
> column.
> AmbiguousTimeError: Cannot infer dst time from Timestamp('2015-11-01
> 01:29:30'), try using the 'ambiguous' argument
> For detailed on repo, see Comment box
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
