[jira] [Commented] (CALCITE-6201) Merge Aggregate node if count distinct column has been deduplicated
[ https://issues.apache.org/jira/browse/CALCITE-6201?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17839705#comment-17839705 ] Sergey Nuyanzin commented on CALCITE-6201: -- added fix version as it suggested to have it even for duplicates https://calcite.apache.org/docs/howto.html#making-a-snapshot > Merge Aggregate node if count distinct column has been deduplicated > --- > > Key: CALCITE-6201 > URL: https://issues.apache.org/jira/browse/CALCITE-6201 > Project: Calcite > Issue Type: Improvement > Components: core >Reporter: Jiajun Xie >Assignee: Jiajun Xie >Priority: Minor > Fix For: 1.37.0 > > > distinct column is redundant because count has distinct. > {code:java} > select count(distinct x) cnt > from( >select distinct sal x from emp > ) t {code} > PlanBefore is > {noformat} > LogicalAggregate(group=[{}], CNT=[COUNT(DISTINCT $0)]) > LogicalAggregate(group=[{5}]) > LogicalTableScan(table=[[CATALOG, SALES, EMP]]){noformat} > PlanAfter should be > {noformat} > LogicalAggregate(group=[{}], CNT=[COUNT(DISTINCT $5)]) > LogicalTableScan(table=[[CATALOG, SALES, EMP]]){noformat} > But `CoreRules.AGGREGATE_MERGE` not support it, we can improve it. > -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Commented] (CALCITE-6201) Merge Aggregate node if count distinct column has been deduplicated
[ https://issues.apache.org/jira/browse/CALCITE-6201?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17809092#comment-17809092 ] Jiajun Xie commented on CALCITE-6201: - Remove `DISTINCT` in count is a better solution. So I create https://issues.apache.org/jira/browse/CALCITE-6214 > Merge Aggregate node if count distinct column has been deduplicated > --- > > Key: CALCITE-6201 > URL: https://issues.apache.org/jira/browse/CALCITE-6201 > Project: Calcite > Issue Type: Improvement > Components: core >Reporter: Jiajun Xie >Assignee: Jiajun Xie >Priority: Minor > > distinct column is redundant because count has distinct. > {code:java} > select count(distinct x) cnt > from( >select distinct sal x from emp > ) t {code} > PlanBefore is > {noformat} > LogicalAggregate(group=[{}], CNT=[COUNT(DISTINCT $0)]) > LogicalAggregate(group=[{5}]) > LogicalTableScan(table=[[CATALOG, SALES, EMP]]){noformat} > PlanAfter should be > {noformat} > LogicalAggregate(group=[{}], CNT=[COUNT(DISTINCT $5)]) > LogicalTableScan(table=[[CATALOG, SALES, EMP]]){noformat} > But `CoreRules.AGGREGATE_MERGE` not support it, we can improve it. > -- This message was sent by Atlassian Jira (v8.20.10#820010)