Github user sudheeshkatkam commented on the pull request:
https://github.com/apache/drill/pull/193#issuecomment-152053724
I think I see the source of confusion (sorry); this patch does not address
that query in the JIRA, which is why Jinfeng asked me to change the title in
one of his comments. Regarding that query, DRILL-3921 helps avoids most of the
execution time, but we still incur the planning time. And my initial approach
address this issue but as mentioned above, this is blocked by DRILL-2288 and
other things.
The new approach actually addresses any query that has a limit 0 above a
blocking operator that consumes all records. And avoiding parallelization made
the query much worse. (Actually, was fast-schema supposed to still kick in? Did
not seem like it from my experiments.)
I tested against a query like `SELECT * FROM (SELECT COUNT(DISTINCT a),
COUNT(DISTINCT b), COUNT(DISTINCT c) FROM very_large_table) T LIMIT 0` and this
completed two orders of magnitude faster.
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
