Re: Optimizing nbtree ScalarArrayOp execution, allowing multi-column ordered scans, skip scan

2024-04-18 Thread Donghang Lin
trategy. That doesn't work if the strategy is invalid. Attached a demo fix. Regards, Donghang Lin (ServiceNow) demo-fix.patch Description: Binary data

Bad estimation for NOT IN clause with big null fraction

2024-04-05 Thread Donghang Lin
for each s2 and subtract it once at the end of the loop as a 2nd heuristic? We then maybe prefer s1disjoint2 over s1disjoint and then s1? Donghang Lin (ServiceNow) [1] https://www.postgresql.org/message-id/flat/CA%2Bmi_8aPEAzBgWZpNTABGM%3DcSq7mRMyPWbMsU8eGmUfH75OTLA%40mail.gmail.com [2] https

Re: Parallel Bitmap Heap Scan reports per-worker stats in EXPLAIN ANALYZE

2024-03-27 Thread Donghang Lin
previous scan stats in workers. Attach v5 with v4 plus the fix to aggregate the counts. Regards, Donghang Lin (ServiceNow) v5-0001-Parallel-Bitmap-Heap-Scan-reports-per-worker-stat.patch Description: Binary data

Re: Parallel Bitmap Heap Scan reports per-worker stats in EXPLAIN ANALYZE

2024-03-25 Thread Donghang Lin
o print worker stats. As mentioned above, they are memorize, sort, and hashaggregate nodes. By the way, is that also a problem for these nodes? > BTW doesn't this also suggest some of the code added to explain.c may > not be quite necessary? Wouldn't it be enough to just "extend" th

Re: interval_ops shall stop using btequalimage (deduplication)

2023-10-17 Thread Donghang Lin
as it's testing btequalimage, but I'll defer it for now. Wanted to know if the above idea sounds sane first... Donghang Lin (ServiceNow) On Thu, Oct 12, 2023 at 4:22 PM Peter Geoghegan wrote: > On Thu, Oct 12, 2023 at 4:10 PM Noah Misch wrote: > > > exactly one case like that post-fi

Re: 2023-08-10 release announcement draft

2023-08-10 Thread Donghang Lin
ual behavior before the fix. What the fix does is to bring exceptions back to the c++ world when OOM happens outside of LLVM. Thanks, Donghang Lin [1] https://github.com/postgres/postgres/blame/accf4f84887eb8b53978a0dbf9cb5656e1779fcb/doc/src/sgml/release-15.sgml#L488-L509 On Wed, Aug 9, 2023 at 7:10

Re: Bug: pg_regress makefile does not always copy refint.so

2022-10-18 Thread Donghang Lin
g you have to run `make` twice to ensure newly built refint.so is copied to the build dir. Regards, Donghang Lin (ServiceNow) From: Alvaro Herrera Date: Friday, October 14, 2022 at 8:15 PM To: Donghang Lin Cc: pgsql-hackers@lists.postgresql.org Subject: Re: Bug: pg_regress makefile does not al

Bug: pg_regress makefile does not always copy refint.so

2022-10-14 Thread Donghang Lin
/regress 4. make We’ll find refint.so is rebuilt in contrib/spi, but not copied over to regress folder. While autoinc.so is rebuilt and copied over. Attach the potential patch to fix the issue. Regards, Donghang Lin (ServiceNow) fix-pgregress-makefile.patch Description: fix-pgregress