pá 17. 2. 2023 v 21:43 odesílatel Jeff Davis <pg...@j-davis.com> napsal:
> On Fri, 2023-02-17 at 18:27 +0100, Pavel Stehule wrote: > > Today I tested icu for Czech sorting. It is a little bit slower, but > > not too much, but it produces partially different results. > > Thank you for trying it. > > If it's a significant slowdown, can you please send more information? > ICU version, libc version, and testcase? > no - this slowdown is not significant - although 1% can looks too much - but it is just two ms It looks so libicu has little bit more expensive initialization, but the execution is little bit faster But when I try to repeat the measurements, the results are very unstable on my desktop :-/ SELECT * FROM obce ORDER BY nazev LIMIT 10 // is faster with glibc little bit SELECT * FROM obce ORDER BY nazev // is faster with libicu You can download dataset https://pgsql.cz/files/obce.sql It is table of municipalities in czech republic (real names) - about 6000 rows I use fedora 37 - so libicu 71.1, glibc 2.36 Regards Pavel > > > select row_number() over (order by nazev collate "cs-x-icu"), nazev > > from obce > > except select row_number() over (order by nazev collate "cs_CZ"), > > nazev from obce; > > > > returns a not empty set. So minimally for Czech collate, an index > > rebuild is necessary > > Yes, that's true of any locale change, provider change, or even > provider version change. > > > -- > Jeff Davis > PostgreSQL Contributor Team - AWS > > >
