[
https://issues.apache.org/jira/browse/CALCITE-512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14239878#comment-14239878
]
Julian Hyde commented on CALCITE-512:
-------------------------------------
I see you have assigned this to yourself. If you want to do this, you should
start from my branch
https://github.com/julianhyde/incubator-calcite/commit/f49276b6bab3115945f0952c4121c2dbd6da0a9e.
Or assign it back to me.
The parsing and validation is done; we just need to translate to an expression
involving indicator columns. {{SqlToRelConverter.lookupAggregates}} is probably
where the work needs to be done. Run JdbcTest.testRunAgg (which calls agg.oq)
and make sure the results look good.
{{GROUP_ID}} and {{GROUPING_ID}} should probably return a BIGINT, not INTEGER
as at present. That will allow for 63 grouping expressions rather than 31.
What do you think should be the behavior if an argument to {{GROUPING_ID}}
occurs multiple times, or if arguments are not in the same order as in the
{{GROUP BY}} clause? I think we already handle the case where an argument to
{{GROUPING_ID}} is not being grouped.
> Add GROUP_ID and GROUPING_ID functions
> --------------------------------------
>
> Key: CALCITE-512
> URL: https://issues.apache.org/jira/browse/CALCITE-512
> Project: Calcite
> Issue Type: Bug
> Reporter: Julian Hyde
> Assignee: Jesus Camacho Rodriguez
>
> Implement GROUP_ID and GROUPING_ID(expr [, expr]...) functions, as they are
> defined in Oracle.
> We already have, as of CALCITE-370, the GROUPING(expr) function. Now define
> GROUPING_ID:
> GROUPING_ID(e2, e1, e0)
> == GROUPING(e2) * 4
> + GROUPING(e1) * 2
> + GROUPING(e0)
> and similarly for different numbers of arguments.
> GROUP_ID() is equivalent to GROUPING_ID(x, y, z), where x, y, z are the
> expressions being grouped.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
