Github user yhuai commented on the pull request:
https://github.com/apache/spark/pull/6682#issuecomment-109790105
Thanks for the example.
For your query, if you write it as
```
select a.value,b.value,c.value,d.value,e.value from
a join b
on a.key = b.key
join c
on a.key = c.key
join d
on a.key = d.key
join e
on a.key = e.key
```
You will find those unnecessary exchange operators are gone (I am just
pointing out a workaround. I am not questioning the issue at all.).
I mean we really need have the notion that all of `key` columns of tables
in the above example are equivalent. I am not sure `meetPartitions` is the
clean solution for this problem.
For the case of full outer joins, I guess you meant all records from all
tables with a null key are shuffled to the same reducer, right? btw, what is
the plan generated by Hive?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
