[
https://issues.apache.org/jira/browse/SPARK-19087?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sean Owen updated SPARK-19087:
------------------------------
Issue Type: Improvement (was: Bug)
I imagine that's on purpose -- the UDF should convert.
> Numpy types fail to be casted to any other types
> ------------------------------------------------
>
> Key: SPARK-19087
> URL: https://issues.apache.org/jira/browse/SPARK-19087
> Project: Spark
> Issue Type: Improvement
> Components: PySpark
> Affects Versions: 1.6.0
> Environment: Red Hat 6.6, Spark 1.6.0, Python 2.7.10 (anaconda)
> Reporter: Ivan SPM
> Priority: Minor
> Labels: numpy, type-converter
>
> An UDF cannot return a numpy type, it has to be one of the Python basic
> types. The error when a numpy type is returned is
> TypeError: ufunc 'nan' not supported for the input types, and the inputs
> could not be safely coerced to any supported types according to the casting
> rule ''safe''
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
