[jira] [Commented] (DRILL-1829) selecting a wide column returns an empty result

["Jason Altekruse (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/DRILL-1829?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14238781#comment-14238781 ] 

Jason Altekruse commented on DRILL-1829:
----------------------------------------

Drill does not currently re-allocate if we cannot write a single value in the pre-allocated memory for the first batch. We will need to detect cases where no values were returned, but the file did contain data and appropriately re-allocate to fit the needs of the wider values.

> selecting a wide column returns an empty result
> -----------------------------------------------
>
>                 Key: DRILL-1829
>                 URL: https://issues.apache.org/jira/browse/DRILL-1829
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Execution - Data Types
>            Reporter: Rahul Challapalli
>         Attachments: 10000.tbl, 100000.tbl
>
>
> git.commit.id.abbrev=142e577
> I attached 2 files. Both of them have just 1 column. The length of the column in one file is 10000 and in the other it is 100000.
> The below queries fail to produce any output
> {code}
> 0: jdbc:drill:schema=dfs.drillTestDir> select columns[0] from `data-shapes/100000.tbl`;
> +--+
> |  |
> +--+
> +--+
> No rows selected (0.09 seconds)
> 0: jdbc:drill:schema=dfs.drillTestDir> select count(*) from `data-shapes/100000.tbl`;
> +--+
> |  |
> +--+
> +--+
> No rows selected (0.092 seconds)
> {code}
> The same queries work when the column width is 10000.
> Let me know if you need anything.



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