[
https://issues.apache.org/jira/browse/IMPALA-5872?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16726978#comment-16726978
]
Paul Rogers commented on IMPALA-5872:
-------------------------------------
The test case info can be used to solve two distinct kinds of problems:
* Planner issues
* Execution issues
For planner issues, we actually don't need an "real" HMS and HDFS with real
data. All Planner access is via an cache in the Coordinator. So, we only need
to populate that in-memory cache.
For execution, we'd need actual data, which is a separate discussion.
> Implement a SQL test case builder for gathering query diagnostics
> -----------------------------------------------------------------
>
> Key: IMPALA-5872
> URL: https://issues.apache.org/jira/browse/IMPALA-5872
> Project: IMPALA
> Issue Type: Improvement
> Components: Frontend
> Reporter: bharath v
> Assignee: bharath v
> Priority: Major
> Labels: supportability
>
> The idea is to implement a test case builder for collecting enough
> diagnostics for a given SQL query so that it can be reproduced/recreated on
> another cluster easily.
> Input: A valid SQL query
> Expected output: A .sql file that can be run using {{impala-shell -f}} and
> has all the schema/table/view definitions required to run the query on a
> target cluster.
> An example:
> {noformat}
> EXPORT TESTCASE INTO OUTFILE 'file:///tmp/repro.sql' select * from view
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
