Github user eychih commented on the issue:
https://github.com/apache/carbondata/pull/2335
In addition, for consistency, please use the following way to remove
PreAggFunction:
val planToRegister =
sparkSession.sql(updatedQuery).queryExecution.analyzed.
transform {
case p @ logical.Project(head :: tail, _) if
head.isInstanceOf[Alias] && head.name.equals("preAgg") => p.copy(projectList =
tail)
case a @ logical.Aggregate(_, head :: tail, _) if
head.isInstanceOf[Alias] && head.name.equals("preAgg") =>
a.copy(aggregateExpressions = tail)
}
---