Github user jinfengni commented on the pull request: https://github.com/apache/drill/pull/397#issuecomment-198407062 Looks good to me. +1 Some additional info 1. This PR has added an option to enable/disable the type inference feature (by default, it's enabled). 2. Internal performance shows this type inference feature, combined with "limit 0" optimization, shows huge improvement for "limit 0" query, some time with orders of magnitude. 3. It has passed the pre-commit test with option enabled/disabled. 4. It has addressed issues found by MapR QA folks during their extensive testing. 5. @hsuanyi will add a doc to the JIRA, showing the scope where the type inference feature would or would not be applied. 6. As a side note, this PR requires one patch in Drill forked Calcite. The change is needed because of the star column changes specifically made for Drill in forked Calcite. For now, it's not feasible to push this change to Calcite master, until the rebasing work is done. @jacques-n , @amansinha100 , @sudheeshkatkam , if you guys have any comment or concern, please let @hsuanyi know. If no concern, I'm going to merge this PR to master branch.
--- 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. ---