[jira] [Resolved] (SPARK-10448) Parquet schema merging should NOT merge UDT

["Hyukjin Kwon (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SPARK-10448?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Hyukjin Kwon resolved SPARK-10448.
----------------------------------
    Resolution: Incomplete

> Parquet schema merging should NOT merge UDT
> -------------------------------------------
>
>                 Key: SPARK-10448
>                 URL: https://issues.apache.org/jira/browse/SPARK-10448
>             Project: Spark
>          Issue Type: Bug
>          Components: SQL
>    Affects Versions: 1.3.1, 1.4.1, 1.5.0
>            Reporter: Cheng Lian
>            Priority: Major
>              Labels: bulk-closed
>
> For example, we may have a UDT {{U}} that maps to a Catalyst {{StructType}} with two fields {{a}} and {{b}}. Later on, we updated {{U}} to {{U'}} by removing {{a}} and adding {{c}}. In this case, Parquet schema merging will give a {{StructType}} with all three fields. But such a {{StructType}} can be mapped to neither {{U}} nor {{U'}}.
> We probably shouldn't allow schema merging over UDT types.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org