Uploaded image for project: 'Calcite'
  1. Calcite
  2. CALCITE-3466

JDBC adapter incorrectly drops GROUP BY clause of sub-query

VotersWatch issueWatchersLinkCloneUpdate Comment AuthorReplace String in CommentUpdate Comment VisibilityDelete Comments
    XMLWordPrintableJSON

Details

    Description

      While convert RelNode to SqlNode, the "group by" statement in subquery was dropped unexpectedly. For example,
      original sql is:

      select a, avg(c1) from 
      (
          select a,sum(d),b as c1 from t group by b,a
      ) as t 
      group by a
      

      RelNode converted by SqlNode is

      LogicalAggregate(group=[{0}], EXPR$1=[AVG($1)])
        LogicalProject(A=[$1], C1=[$0])
          LogicalAggregate(group=[{0, 1}], EXPR$1=[SUM($2)])
            LogicalProject(B=[$1], A=[$0], D=[$3])
              EnumerableTableScan(table=[[T_TESTPROJECTMERGE]])
      

      And then convert this RelNode to SqlNode, the sql of SqlNode is like this:

      SELECT `A`, AVG(`B`)
      FROM `T`
      GROUP BY `A`
      

      We can find that "group by" in subquery was dropped which may cause SqlNode converted by RelNode being not equivalent to RelNode.

      After debugging, I've found code
      at https://github.com/apache/calcite/blob/aa98c1575762718b238af8ddd44a390607de7602/core/src/main/java/org/apache/calcite/rel/rel2sql/RelToSqlConverter.java#L333
      calls org.apache.calcite.rel.rel2sql.SqlImplementor.Builder#setGroupBy without judge whether "group by" already exists.
      Finally, I think in this case, the main select and sub-select shouldn't be merged.

      Attachments

        Activity

          This comment will be Viewable by All Users Viewable by All Users
          Cancel

          People

            wangweidong Wang Weidong
            wangweidong Wang Weidong
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Time Tracking

                Estimated:
                Original Estimate - Not Specified
                Not Specified
                Remaining:
                Remaining Estimate - 0h
                0h
                Logged:
                Time Spent - 8h 40m
                8h 40m

                Slack

                  Issue deployment