Github user kumarvishal09 commented on a diff in the pull request:
https://github.com/apache/carbondata/pull/1724#discussion_r158718262
--- Diff:
integration/spark2/src/main/scala/org/apache/spark/sql/hive/CarbonPreAggregateRules.scala
---
@@ -1264,13 +1264,16 @@ object CarbonPreAggregateDataLoadingRules extends
Rule[LogicalPlan] {
override def apply(plan: LogicalPlan): LogicalPlan = {
val validExpressionsMap =
scala.collection.mutable.LinkedHashMap.empty[String, NamedExpression]
plan transform {
- case aggregate@Aggregate(_, aExp, _) if
validateAggregateExpressions(aExp) =>
+ case aggregate@Aggregate(groupingExpressions, aExp, _) if
validateAggregateExpressions(aExp) =>
aExp.foreach {
case alias: Alias =>
validExpressionsMap ++=
validateAggregateFunctionAndGetAlias(alias)
case _: UnresolvedAlias =>
case namedExpr: NamedExpression =>
validExpressionsMap.put(namedExpr.name, namedExpr)
}
+ groupingExpressions foreach {
--- End diff --
I don't get the purpose of this change data loading rules is only to update
the average column as average columns data will be stored in two column(sum,
count), can u please explain why we need this change??
---
