Re: [pg_trgm] Making similarity(?, ?) < ? use an index

Поиск
Список
Период
Сортировка
От Greg Navis
Тема Re: [pg_trgm] Making similarity(?, ?) < ? use an index
Дата
Msg-id CAA6WWt-mEHVnDxn=gLh+dmTVSfRgtULjMand-kHq8uq7eTi2Bw@mail.gmail.com
обсуждение исходный текст
Ответ на Re: [pg_trgm] Making similarity(?, ?) < ? use an index  (Artur Zakirov <a.zakirov@postgrespro.ru>)
Список pgsql-general
Artur, thanks for help. I managed to add the new strategy to the index. Hurray! I also discovered a bug in the process that I reported via the form.

I still have a few questions:

1. Naming - pg_trgm_match, match, threshold, trgm_check_match, ThresholdStrategyNumber - are these good names?
2. I made trgm_check_match IMMUTABLE. Are there any other modifies that should be there?
3. I defined % (text, pg_trgm_match) but didn't provide a commutator and other helper procedures. Which of them should I implement?
4. Can I obtain query and nlimit with less code?
5. The attached patch replaced "res = (*(int *) &tmpsml == *(int *) &nlimit || tmpsml > nlimit);" with "res = (tmpsml >= nlimit);" to fix the bug on my machine. I'm not sure whether that's the long-term fix we want to have. It's just there to help me make progress with trigrams.

Thanks for help.

Cheers
Greg
Вложения

В списке pgsql-general по дате отправления:

Предыдущее
От: Adarsh Sharma
Дата:
Сообщение: Re: R: Vacuum full: alternatives?
Следующее
От: Martín Marqués
Дата:
Сообщение: Re: R: Vacuum full: alternatives?