[
https://issues.apache.org/jira/browse/SPARK-59721?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Anurag Mantripragada updated SPARK-59721:
-----------------------------------------
Description:
When a DSv2 scan reports a KeyGroupedPartitioning or an ordering that
references a column the relation can't resolve, the query fails with Unable to
resolve X given [...]. V2ScanPartitioningAndOrdering should check the columns
referenced by the reported partition keys and ordering (recursively, including
nested transform arguments) against the relation before converting them. If any
can't be resolved, including a missing nested field or an ambiguous reference,
it should log a warning naming the relation, the scan class, and the columns,
and ignore the report: the partitioning falls back to unknown partitioning, and
the ordering is dropped.
Found while reviewing SPARK-58111 (PR
[#55518|https://github.com/apache/spark/pull/55518]); this bug is general and
pre-existing, not introduced there.
was:
V2ExpressionUtils.toCatalystOpt and toCatalystTransformOpt should gracefully
return None when a data source's reported partition/bucket/transform key cannot
be resolved against the relation. This is similar to
V2ScanPartitioningAndOrdering's partitioning() rule (SPARK-59248) relies on
exactly this to degrade a KeyGroupedPartitioning to unknown partitioning
instead of failing the query when a key can't be resolved.
Found while reviewing SPARK-58111 (PR
[#55518|https://github.com/apache/spark/pull/55518]); this bug is general and
pre-existing, not
introduced there.
> Fall back to unknown partitioning instead of failing when a reported
> partition key cannot be resolved
> -----------------------------------------------------------------------------------------------------
>
> Key: SPARK-59721
> URL: https://issues.apache.org/jira/browse/SPARK-59721
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 5.0.0
> Reporter: Anurag Mantripragada
> Priority: Major
> Labels: pull-request-available
>
> When a DSv2 scan reports a KeyGroupedPartitioning or an ordering that
> references a column the relation can't resolve, the query fails with Unable
> to resolve X given [...]. V2ScanPartitioningAndOrdering should check the
> columns referenced by the reported partition keys and ordering (recursively,
> including nested transform arguments) against the relation before converting
> them. If any can't be resolved, including a missing nested field or an
> ambiguous reference, it should log a warning naming the relation, the scan
> class, and the columns, and ignore the report: the partitioning falls back to
> unknown partitioning, and the ordering is dropped.
> Found while reviewing SPARK-58111 (PR
> [#55518|https://github.com/apache/spark/pull/55518]); this bug is general and
> pre-existing, not introduced there.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
