You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@drill.apache.org by "Jacques Nadeau (JIRA)" <ji...@apache.org> on 2016/06/27 04:05:51 UTC

[jira] [Commented] (DRILL-4754) Missing values are not missing

    [ https://issues.apache.org/jira/browse/DRILL-4754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15350394#comment-15350394 ] 

Jacques Nadeau commented on DRILL-4754:
---------------------------------------

This is a current known limitation of Drill 1.6/current Drill master. We can't distinguish between empty list and null. (The vectors don't currently allow it.) Same goes for treating empty map and null map the same.

I believe there are already some JIRAs outstanding that are specifically focused on this issue. (e.g. support nullable maps and nullable lists)

> Missing values are not missing
> ------------------------------
>
>                 Key: DRILL-4754
>                 URL: https://issues.apache.org/jira/browse/DRILL-4754
>             Project: Apache Drill
>          Issue Type: Bug
>            Reporter: Ted Dunning
>
> If I have a query which reads from a JSON file where a field is a list or is missing, then the records where the field should missing will instead have a value for that field that is an empty list:
> {code}
> 0: jdbc:drill:> select * from maprfs.ted.`bug.json`;
> +----+--------+------+
> | *a*  |   *b*    |  *c*   |
> | 3  | [3,2]  | xyz  |
> | 7  | []     | wxy  |
> | 7  | []     | null  |
> +----+--------+------+
> 2 rows selected (1.279 seconds)
> {code}
> where the file in question contains these three records:
> {code}
> {'a':3, 'b':[3,2], 'c':'xyz'}
> {'a':7, 'c':'wxy'}
> {"a":7, "b":[]}
> {code}
> The problem is in the second record of the result. I would have expected b to have had the value NULL.
> I am using drill-1.6.0.



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