[
https://issues.apache.org/jira/browse/DRILL-8501?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17863534#comment-17863534
]
ASF GitHub Bot commented on DRILL-8501:
---------------------------------------
cgivre commented on PR #2921:
URL: https://github.com/apache/drill/pull/2921#issuecomment-2212305704
> Before I approve, did you consider making these JSON parsing settings
parameters of the function itself? It feels odd to me that `store.json.*`
settings could influence UDFs too. I'm not sure why they aren't storage plugin
config, rather than global config, in the first place...
@jnturton I thought about doing exactly what you're describing. Here's the
thing. We started some work a while ago to get rid of all the non-EVF2 readers
in Drill. It turns out that there are a few places which still use the old
non-EVF JSON reader. Specifically, this UDF, the Druid Storage Plugin and the
MongoDB storage plugin. I started work on
[Drill-8316](https://github.com/apache/drill/pull/2657) which addresses the
Druid plugin and [Drill-8329](https://github.com/apache/drill/pull/2567)
addresses converting the UDF. Neither one of these were a high priority so
they're kind of sitting at the moment.
I agree with your premise that the whole idea of having global settings for
file formats (including parquet) is not the best idea.
> Json Conversion UDF Not Respecting System JSON Options
> ------------------------------------------------------
>
> Key: DRILL-8501
> URL: https://issues.apache.org/jira/browse/DRILL-8501
> Project: Apache Drill
> Issue Type: Bug
> Components: Storage - JSON
> Affects Versions: 1.21.2
> Reporter: Charles Givre
> Assignee: Charles Givre
> Priority: Major
> Fix For: 1.22.0
>
>
> The convert_fromJSON() UDF does not respect the system JSON options of
> allTextMode and readAllNumbersAsDouble.
> This PR fixes that.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
