[
https://issues.apache.org/jira/browse/HIVE-16034?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15883318#comment-15883318
]
Ashutosh Chauhan commented on HIVE-16034:
-----------------------------------------
This can be made more robust by making sure type of dimensions is in
PrimitiveGrouping.STRING_GROUP and type of metrics is in
PrimitiveGrouping.NUMERIC_GROUP and throw exception if this is not the case.
Outside of these groupings we cant handle other types any way.
Current assumption of default assuming dimensions make code less predictable
and results in hard to debug issues.
> Hive/Druid integration: Fix type inference for Decimal DruidOutputFormat
> ------------------------------------------------------------------------
>
> Key: HIVE-16034
> URL: https://issues.apache.org/jira/browse/HIVE-16034
> Project: Hive
> Issue Type: Bug
> Components: Druid integration
> Affects Versions: 2.2.0
> Reporter: Jesus Camacho Rodriguez
> Assignee: Jesus Camacho Rodriguez
> Attachments: HIVE-16034.patch
>
>
> We are extracting the type name by String, which might cause issues, e.g.,
> for Decimal, where type includes precision and scale. Instead, we should
> check the PrimitiveCategory enum.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
