[
https://issues.apache.org/jira/browse/SPARK-53793?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Yuchuan Huang updated SPARK-53793:
----------------------------------
Description:
This ticket proposes to add a utility class to enable the evaluation of an
InternalRow using a DSv2 predicate. This would be helpful for partition
pruning, where the [runtime filters are DSv2
predicates|https://github.com/apache/spark/blob/master/sql/catalyst/src/main/java/org/apache/spark/sql/connector/read/SupportsRuntimeV2Filtering.java#L66]
and the partitionValue are internalRows (for [partitionFiles in
Spark|https://github.com/apache/spark/blob/65ff85a31fe8a8ea4a2ba713ba2c624709ce815a/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/FileScanRDD.scala#L58]).
In this way, partitionFiles can be pruned directly with DSv2 predicates at the
scan level.
To enable this, a DSv2 predicate will be converted to a catalyst expression,
and then create an evaluator.
> Use DSv2 predicate to evaluate InternalRow
> ------------------------------------------
>
> Key: SPARK-53793
> URL: https://issues.apache.org/jira/browse/SPARK-53793
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 4.1.0
> Reporter: Yuchuan Huang
> Priority: Major
>
> This ticket proposes to add a utility class to enable the evaluation of an
> InternalRow using a DSv2 predicate. This would be helpful for partition
> pruning, where the [runtime filters are DSv2
> predicates|https://github.com/apache/spark/blob/master/sql/catalyst/src/main/java/org/apache/spark/sql/connector/read/SupportsRuntimeV2Filtering.java#L66]
> and the partitionValue are internalRows (for [partitionFiles in
> Spark|https://github.com/apache/spark/blob/65ff85a31fe8a8ea4a2ba713ba2c624709ce815a/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/FileScanRDD.scala#L58]).
> In this way, partitionFiles can be pruned directly with DSv2 predicates at
> the scan level.
>
> To enable this, a DSv2 predicate will be converted to a catalyst expression,
> and then create an evaluator.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
