[jira] [Resolved] (SPARK-39806) Queries accessing METADATA struct crash on partitioned tables

["Wenchen Fan (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SPARK-39806?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Wenchen Fan resolved SPARK-39806.
---------------------------------
    Fix Version/s: 3.3.1
                   3.4.0
       Resolution: Fixed

Issue resolved by pull request 37214
[https://github.com/apache/spark/pull/37214]

> Queries accessing METADATA struct crash on partitioned tables
> -------------------------------------------------------------
>
>                 Key: SPARK-39806
>                 URL: https://issues.apache.org/jira/browse/SPARK-39806
>             Project: Spark
>          Issue Type: Improvement
>          Components: SQL
>    Affects Versions: 3.3.0
>            Reporter: Ala Luszczak
>            Assignee: Ala Luszczak
>            Priority: Major
>             Fix For: 3.3.1, 3.4.0
>
>
> There is a problem with a projection we use in `FileScanRDD` to join the metadata row to the row produced by the reader.
> https://github.com/apache/spark/blob/e4ca8424474e571d8e137388fe5d54732b68c2f3/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/FileScanRDD.scala#L128-L133
> The issue is that the projection omits partition columns. As a result, the expressions down the line return a malformed row. The errors crash the query, but the exact message can vary (for example: failed assertion on number of fields in the row, accessing field of incorrect type).
> This defect affects only readers producing rows (as opposed to batches), and only data sets using dynamic partitioning.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

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