[
https://issues.apache.org/jira/browse/KYLIN-5713?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17786625#comment-17786625
]
Hongrong Cao commented on KYLIN-5713:
-------------------------------------
Dev Design
The reason is that calcite does not support this function.
org.apache.kylin.query.engine.ProjectSchemaFactory#addUDFs to register the
function into calcite, so that calcite recognizes this function will not report
errors, because spark supports this function, do not need to write the
implementation of the udf way, just need to throw CalciteNotSupportExcetion to
make the sql routed to spark execution.
> Queries supporting regexp_extract function点击并应用
> -----------------------------------------------
>
> Key: KYLIN-5713
> URL: https://issues.apache.org/jira/browse/KYLIN-5713
> Project: Kylin
> Issue Type: Improvement
> Affects Versions: 5.0-beta
> Reporter: Hongrong Cao
> Assignee: Ruixuan Zhang
> Priority: Minor
> Fix For: 5.0.0
>
>
> regexp_extract is supported in spark-sql, but not in kylin.
> In kylin, you can set up regexp_extract related computable columns, verify
> that they work, and that they can be queried and decompressed normally after
> they are built.
> Querying the regexp_extract function alone will also indicate that it is not
> supported, so it should be due to the fact that calcite does not support this
> function.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
