apache / drill

Apache Drill is a distributed MPP query layer for self describing data

https://drill.apache.org/

Apache License 2.0

1.95k stars 979 forks source link

DRILL-8506: Ignore JSON Elements with Empty Keys #2935

Closed cgivre closed 2 months ago

cgivre commented 3 months ago

DRILL-8506: Ignore JSON Elements with Empty Keys

Description

If Drill encounters a JSON field with an empty key, currently Drill throws an exception. This behavior prevents Drill from reading the rest of the data. This PR modifies Drill's behavior by simply ignoring JSON data with empty keys. Drill will generate a log warning message when this happens.

Documentation

No significant user facing changes.

Testing

Ran existing unit tests.

cgivre commented 2 months ago

Unless anyone has any objections, I'm going to merge this. It seems that whatever was going wrong with the CI has magically fixed itself. Although some unrelated Hive tests are failing due to connections being dropped, but only for Hadoop 2....

martingaleano commented 1 month ago

Hi @cgivre , thanks for adding this change! I really need it. I've noticed that this is not included in the docker versions yet (I mean, there isn't any release), do you know when I can run the docker image with your changes?

jnturton commented 1 month ago

@martingaleano we publish Docker images that track the Drill master branch. Anything tagged with "master" here.