[
https://issues.apache.org/jira/browse/HIVE-24596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17293773#comment-17293773
]
Zoltan Haindrich commented on HIVE-24596:
-----------------------------------------
Yes, reducing the round-trips in problematic cases would be very usefull!
I'm still a bit confused; based on the description I've saw so far I (somehow)
was expecting sql statements as output...and I kinda still feel like that
should be the case.
note that in the PR 'explain ddl' outputs seem more like standard explains...so
I feel like I'm missing basic with the concept.
Maybe it will be easier to understand your idea thru an example:
{code}
$ create table t (a integer);
$ create table t2 (a integer);
$ explain ddl select 1 from t join t2 where t.a=t2.a;
{code}
for the above I would expect to see a "create table t (a integer)" in the
output...could you give a theoretical transcript?
> Explain ddl for debugging
> -------------------------
>
> Key: HIVE-24596
> URL: https://issues.apache.org/jira/browse/HIVE-24596
> Project: Hive
> Issue Type: Improvement
> Reporter: Rajesh Balamohan
> Assignee: Harshit Gupta
> Priority: Major
> Labels: pull-request-available
> Time Spent: 20m
> Remaining Estimate: 0h
>
> For debugging query issues, basic details like table schema, statistics,
> partition details, query plans are needed.
> It would be good to have "explain ddl" support, which can generate these
> details. This can help in recreating the schema and planner issues without
> sample data.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
