You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@drill.apache.org by "Adam Gilmore (JIRA)" <ji...@apache.org> on 2015/04/08 09:38:12 UTC

[jira] [Updated] (DRILL-2361) Column aliases cannot include dots

     [ https://issues.apache.org/jira/browse/DRILL-2361?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Adam Gilmore updated DRILL-2361:
--------------------------------
    Attachment: DRILL-2361.2.patch.txt

Here is a patch for more of the situations plus tests.

I would really like to understand the point behind the "check" in FieldReference for dots in the names.  I can't see exactly where this serves a purpose and it seems to hinder more cases than help.

> Column aliases cannot include dots
> ----------------------------------
>
>                 Key: DRILL-2361
>                 URL: https://issues.apache.org/jira/browse/DRILL-2361
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Query Planning & Optimization
>    Affects Versions: 0.7.0, 0.8.0
>            Reporter: Adam Gilmore
>            Assignee: Aman Sinha
>            Priority: Minor
>         Attachments: DRILL-2361.1.patch.txt, DRILL-2361.2.patch.txt
>
>
> The Drill query engine rejects column aliases for aggregates or group by fields where the alias includes a dot.
> For example:
> {code}select count(*) as `test.field` from cp.`employee.json`;{code}
> throws an error with "Field references must be singular names."
> Having reviewed the code, I can't quite see a reason for it, and when using FieldReference.getWithQuotedRef instead in AggPrelBase, the issue is alleviated with no visible side effects.
> I'm sure there was a reason for this code in FieldReference, but I can't seem to work out why, and it seems entirely reasonable to allow a dot in a column alias.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)