Xiening Dai

> Quotes Julian's comment in the mailing list: > > > Is there a change we could make to the rule API that would make it easier to customize rules?...

@danny0405 had a good feedback - since ProjectMergeRule can also be used by HepPlanner, and in some cases user would use HepPlanner to merge physical project nodes at the end...

> Have you seen https://issues.apache.org/jira/browse/CALCITE-2223 ? > I think 3455 duplicates 2223, and, unfortunately, there's no trivial solution for it. I don't think this is a duplication. This patch tries...

@vlsi @zabetak @rubenada @hsyuan I've updated this patch based on your feedback. Now I create separate LOGICAL_INSTANCE for rules that potentially would only match for logical nodes. Now Calcite users...

> I have a question related to CALCITE-3455 but not necessarily related to this PR: > How about providing an approach for RelOptRule and allow to specify that rels matched...

Thanks @zabetak for the pointer. After reviewing the whole thread, I now understand why @vlsi believes this is dup of 2223. I think "single convention" and "convention preserving" rules are...

hi @lemire , do you have any pointer to the Parquet team's work on making Java implementation interoperable with c++? I am interested in looking into this problem.

Ok, after some research, it seems to me that the endian is handled by IntBasedBitPackingGenerator.java (https://github.com/apache/parquet-mr/blob/dc61e510126aaa1a95a46fe39bf1529f394147e9/parquet-generator/src/main/java/org/apache/parquet/encoding/bitpacking/IntBasedBitPackingGenerator.java). The class generates codes for both little endian and big endian packing/unpacking methods. Besides...