Github user gatorsmile commented on a diff in the pull request:
https://github.com/apache/spark/pull/14322#discussion_r71965622
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/FileSourceStrategy.scala
---
@@ -135,9 +135,17 @@ private[sql] object FileSourceStrategy extends
Strategy with Logging {
PUSHED_FILTERS -> pushedDownFilters.mkString("[", ", ", "]"),
INPUT_PATHS -> fsRelation.location.paths.mkString(", "))
+ // If the required attributes does not have the partitioning
columns, we do not need
+ // to scan the partitioning columns. If partitioning columns are
selected, the column order
+ // of partitionColumns is fixed in rdd. Thus, we always scan all the
partitioning columns.
+ val scannedColumns = if
(requiredAttributes.intersect(partitionSet).nonEmpty) {
+ readDataColumns ++ partitionColumns
+ } else {
+ readDataColumns
+ }
val scan =
DataSourceScanExec.create(
- readDataColumns ++ partitionColumns,
+ scannedColumns,
--- End diff --
This is solution one that requires few code changes but gets most benefits.
We can have another solution that covers more cases by changing the RDD
generation.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
