Hello, community! I would like to ask for your help with the review of this PR [1].
It’s been a while since I worked on this for the last time, but now I’m really looking forward to driving this till the end. This change addresses the issue described in CALCITE-4913 [2]. Basically, I want the correlates inside the select list to be deduplicated. To make this possible, information about variables set by the current relation should be kept within the Project relation (most changes are related to this part). Besides, I found out that FieldTrimmer doesn’t update any correlates with respect to the trimmed input, thus the correlated variable refers to the wrong data type. So, I’ve addressed this issue too, but only for the projection node (see [3]). About the previous iteration of review: I’ve addressed most of the comments, but the rest I believe could be addressed in the separate tickets (like investigate the rest of the rules to get sure that variables are propagated properly). I really appreciate any help. [1] https://github.com/apache/calcite/pull/2623 [2] https://issues.apache.org/jira/browse/CALCITE-4913 [3] https://github.com/apache/calcite/pull/2623/files#diff-e6bbdf680fd47326b24b179d7ba13e3e48d1056ba93bb0411bae725813563946R536 -- Regards, Konstantin Orlov
