alexeykudinkin commented on code in PR #5244:
URL: https://github.com/apache/hudi/pull/5244#discussion_r846686802
##########
hudi-spark-datasource/hudi-spark-common/src/main/scala/org/apache/hudi/HoodieFileIndex.scala:
##########
@@ -196,12 +191,20 @@ case class HoodieFileIndex(spark: SparkSession,
* @return list of pruned (data-skipped) candidate base-files' names
*/
private def lookupCandidateFilesInMetadataTable(queryFilters:
Seq[Expression]): Try[Option[Set[String]]] = Try {
- if (!isDataSkippingEnabled || queryFilters.isEmpty ||
!HoodieTableMetadataUtil.getCompletedMetadataPartitions(metaClient.getTableConfig)
- .contains(HoodieTableMetadataUtil.PARTITION_NAME_COLUMN_STATS)) {
+ // NOTE: Data Skipping is only effective when it references columns that
are indexed w/in
+ // the Column Stats Index (CSI). Following cases could not be
effectively handled by Data Skipping:
+ // - Expressions on top-level column's fields (ie, for ex filters
like "struct.field > 0", since
+ // CSI only contains stats for top-level columns, in this case
for "struct")
+ // - Any expression not directly referencing top-level column
(for ex, sub-queries, since there's
+ // nothing CSI in particular could be applied for)
+ lazy val queryReferencedColumns = collectReferencedColumns(spark,
queryFilters, schema)
+
+ if (!isMetadataTableEnabled || !isColumnStatsIndexAvailable ||
!isDataSkippingEnabled) {
Review Comment:
@nsivabalan i'm addressing this problem in a separate PR to avoid
overloading this one: #5275
--
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]
