> > IMO, It seems more readable to extract all the check that we can do > > before the safety-check and put them in the new function. > > > > Please consider it for further review. > > > > I've updated your v2 patches and altered some comments and documentation > changes (but made no code changes) - please compare against your v2 patches, > and see whether you agree with the changes to the wording. > In the documentation, you will also notice that in your V2 patch, it says > that the "parallel_dml_enabled" table option defaults to false. > As it actually defaults to true, I changed that in the documentation too.
Hi greg, Thanks a lot for the document update, LGTM. Attaching v4 patches with the changes: * fix some typos in the code. * store partitioned reloption in a separate struct PartitionedOptions, making it more standard and easier to expand in the future. Please consider it for further review. Best regards, Houzj
v4_0001-guc-option-enable_parallel_dml-src.patch
Description: v4_0001-guc-option-enable_parallel_dml-src.patch
v4_0002-guc-option-enable_parallel_dml-doc-and-test.patch
Description: v4_0002-guc-option-enable_parallel_dml-doc-and-test.patch
v4_0003-reloption-parallel_dml-src.patch.patch
Description: v4_0003-reloption-parallel_dml-src.patch.patch
v4_0004-reloption-parallel_dml-test-and-doc.patch
Description: v4_0004-reloption-parallel_dml-test-and-doc.patch
