[
https://issues.apache.org/jira/browse/TAJO-1420?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14395716#comment-14395716
]
ASF GitHub Bot commented on TAJO-1420:
--------------------------------------
Github user jihoonson commented on the pull request:
https://github.com/apache/tajo/pull/439#issuecomment-89574266
Definitely, we should break a big problem into smaller ones if possible and
helpful.
IMO, this is not a case of the above one. You already break the big
problem, that is supporting grouping sets, into small ones including supporting
the syntax. I mean, we should do more important and core work first. I think
that this tradition is commonly shared by many development projects. When we
develop something, we first design the whole architecture, and then break it
into several modules. After that, we implement core modules first because
trivial modules are not anythings without the core ones.
On opening the legacy code, it may be helpful as you said. However, whether
the legacy code exists or not, we should adopt an efficient algorithm. To do
so, we should investigate many other projects and papers, and then discuss
them. This is one of the main reasons of using Jira.
Anyway, the performance of legacy code is not acceptable for practical
workloads. So, it wouldn't be helpful so much.
Finally, you can reuse the issue if you want. Please update the issue title
and description to be appropriate for the changed issue.
> Support GROUPING SET syntax
> ---------------------------
>
> Key: TAJO-1420
> URL: https://issues.apache.org/jira/browse/TAJO-1420
> Project: Tajo
> Issue Type: Sub-task
> Components: parser
> Affects Versions: 0.10.0
> Reporter: Dongjoon Hyun
> Assignee: Dongjoon Hyun
> Priority: Minor
> Fix For: 0.10.1
>
> Attachments: TAJO-1420.patch
>
>
> This issue is about SQL parsing only. As of Tajo 0.10.0,
> {code:sql}
> default> create table t(a int, b int);
> OK
> default> select * from t group by cube(a,b);
> ERROR: Cube is not supported yet
> default> select * from t group by rollup(a,b);
> ERROR: Rollup is not supported yet
> default> select * from t group by grouping set(a,b);
> ERROR: syntax error at or near 'grouping'
> {code}
> The last error should be like the following.
> {code:sql}
> default> select * from t group by grouping set(a,b);
> ERROR: GroupingSet is not supported yet
> {code}
> As you know,
> {code:sql}
> GROUP BY a, b, c WITH CUBE
> {code}
> is equivalent to
> {code:sql}
> GROUP BY a, b, c GROUPING SETS ( (a, b, c), (a, b), (b, c), (a, c), (a), (b),
> (c), ( ))
> {code}
> And,
> {code:sql}
> GROUP BY a, b, c, WITH ROLLUP
> {code}
> is equivalent to
> {code:sql}
> GROUP BY a, b, c GROUPING SETS ( (a, b, c), (a, b), (a), ( ))
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
