[jira] [Commented] (CALCITE-4342) More aggregate functions should be marked as splittable and duplicate-insensitive

["Julian Hyde (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CALCITE-4342?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17215492#comment-17215492 ] 

Julian Hyde commented on CALCITE-4342:
--------------------------------------

I agree with this, but let's not introduce the new term 'duplicate-insensitive'. The terminology of the existing API will be sufficient.

While you're there, can you look into {{MaterializedViewAggregateRule.getRollup}} and {{SubstitutionVisitor.getRollup}}? It seems to me that the information about how an aggregate function decomposes/rolls up is spread over several places. Ideally, it should be defined in the aggregate function, not in several rules.

> More aggregate functions should be marked as splittable and duplicate-insensitive
> ---------------------------------------------------------------------------------
>
>                 Key: CALCITE-4342
>                 URL: https://issues.apache.org/jira/browse/CALCITE-4342
>             Project: Calcite
>          Issue Type: Improvement
>          Components: core
>            Reporter: Liya Fan
>            Assignee: Liya Fan
>            Priority: Major
>
> As discussed in the mailing list, there are more aggregate functions (e.g. SINGLE_VALUE, ANY_VALUE) that can be classified as splittable and duplicate-insensitive. We need to tell the optimizer explicitly about this, so more optimizations can be applied in various scenarios. 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)