Github user ravipesala commented on a diff in the pull request:
https://github.com/apache/carbondata/pull/1672#discussion_r157738771
--- Diff:
integration/spark2/src/main/scala/org/apache/spark/sql/execution/strategy/CarbonLateDecodeStrategy.scala
---
@@ -130,6 +130,44 @@ private[sql] class CarbonLateDecodeStrategy extends
SparkStrategy {
table.carbonTable.getTableInfo.serialize())
}
+ /**
+ * Converts to physical RDD of carbon after pushing down applicable
filters.
+ * @param relation
+ * @param projects
+ * @param filterPredicates
+ * @param scanBuilder
+ * @return
+ */
+ private def pruneFilterProject(
+ relation: LogicalRelation,
+ projects: Seq[NamedExpression],
+ filterPredicates: Seq[Expression],
+ scanBuilder: (Seq[Attribute], Array[Filter],
+ ArrayBuffer[AttributeReference], Seq[String]) => RDD[InternalRow])
= {
--- End diff --
Yes, you are right. It is difficult to track the parameters meaning. Even I
also find it difficult pass a new parameter partitions but this is old code
added as part of spark2.0 support and may we can refactor later.
---
