vitaliili-db commented on code in PR #55887:
URL: https://github.com/apache/spark/pull/55887#discussion_r3250285283
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/PushDownUtils.scala:
##########
@@ -187,6 +193,90 @@ object PushDownUtils extends Logging {
}
}
+ /**
+ * Pushes runtime filters into `scan` and re-plans its input partitions. For
scans whose
+ * `outputPartitioning` is a [[KeyedPartitioning]] (SPJ-active), validates
that the data source
+ * preserved the original partitioning and pads with `None` to preserve key
alignment with the
+ * pre-filter partition set.
+ *
+ * Must be called at execute time: runtime filters carry
[[DynamicPruningExpression]] and
+ * scalar-subquery references whose values are only resolved after their
broadcast/subquery
+ * side completes. Callers should wrap the result in a `lazy val` so the
mutating
+ * [[pushRuntimeFilters]] call runs at most once per scan instance.
+ *
+ * @param scan the V2 scan to push filters into
+ * @param runtimeFilters runtime filters to translate and push
+ * @param partitionPredicateSchema by-name schema for iterative
[[PartitionPredicate]] pushdown
+ * @param output scan output attributes
+ * @param outputPartitioning Spark-side output partitioning (used for
SPJ validation)
+ * @param inputPartitions by-name original (unfiltered)
partitions; consulted only when
+ * no runtime filters fire, so callers can
compute it lazily
+ * @return one entry per original input partition: `Some(part)` for
surviving partitions and
+ * `None` for partition keys whose splits were entirely pruned (SPJ
alignment)
+ */
+ def filterAndPlanPartitions(
Review Comment:
this is future work, refactoring to simplify BatchScanExec and centralize
logic
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/PushDownUtils.scala:
##########
@@ -139,12 +141,16 @@ object PushDownUtils extends Logging {
* the first pass are used to derive PartitionPredicates in the second pass,
avoiding duplicate
* pushdown.
*
+ * The partition-predicate schema is passed by-name so callers that cannot
supply one (no
+ * partition transforms available) or whose scan does not opt into iterative
pushdown pay no
+ * derivation cost.
+ *
* @return true if any filters were pushed to the data source
*/
def pushRuntimeFilters(
scan: Scan,
runtimeFilters: Seq[Expression],
- table: Table,
+ partitionPredicateSchema: => Option[Seq[PartitionPredicateField]],
Review Comment:
done
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
