shahidki31 commented on a change in pull request #24715: [SPARK-25474][SQL]
Data source tables support fallback to HDFS for size estimation
URL: https://github.com/apache/spark/pull/24715#discussion_r315120002
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/DataSourceStrategy.scala
##########
@@ -619,3 +619,34 @@ object DataSourceStrategy {
(nonconvertiblePredicates ++ unhandledPredicates, pushedFilters,
handledFilters)
}
}
+
+
+/**
+ * Defines default table statistics if table statistics are not available.
+ */
+class DetermineTableStats(session: SparkSession) extends Rule[LogicalPlan] {
+
+ private val conf = session.sessionState.conf
+
+ override def apply(plan: LogicalPlan): LogicalPlan = plan resolveOperators {
+ // For the data source table, we only recalculate the table statistics
Review comment:
Yes. Both the PR checks, if using `CatalogFileIndex` and table stats is
empty, then fallback to HDFS to get the statistics. And both the PRs never
return statistics from `CatalogFileIndex` irrespective of the stats is empty or
not, if enable fallback. That is why I said the behavior is same.
The pr is looks good. only concern is, if the catalogFileIndex already has
statistics, then in the rules, we shouldn't fallback to HDFS, because that was
the issue with #22502. Thanks.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
