Re: Tsearch Index Size and GiST vs. GIN

Поиск
Список
Период
Сортировка
От Teodor Sigaev
Тема Re: Tsearch Index Size and GiST vs. GIN
Дата
Msg-id 4551A321.9000409@sigaev.ru
обсуждение исходный текст
Ответ на Tsearch Index Size and GiST vs. GIN  (Richard Whidden <richard@armchair.mb.ca>)
Список pgsql-hackers

Richard Whidden wrote:
> Due to popular demand (1 person), I've compared sizes with 90 and 100 
> fillfactors, along with using the new GIN index.

First, GIN doesn't utilize fillfactor option yet.

> 
> Findings were not surprising, except for the GIN indexes, which doubled 
> in size.
Second, In opposite to GiST, size of GIN index strongly depends on document 
collection. Each unique word in collection has list of pointer to document in 
which it occur. Size of pointer is a 6 byte. So word with 4 bytes length will 
occupy at least 4 + N_docs_with_word * 6.

Fortunately, for searching it's needed to read very small part of index, so GIN 
will be faster.


> 
> After several ALTER/RE INDEXes, here they are:
> 
> GiST
> ----
> 8.1 = 94990
> 8.2 FF90 = 106244 relpages (8k)
> 8.2 FF100 = 95049
> 
> GIN
> ---
> FF100 = 197702
> 
> 
> Richard Whidden
> 
> ---------------------------(end of broadcast)---------------------------
> TIP 1: if posting/reading through Usenet, please send an appropriate
>       subscribe-nomail command to majordomo@postgresql.org so that your
>       message can get through to the mailing list cleanly

-- 
Teodor Sigaev                                   E-mail: teodor@sigaev.ru
  WWW: http://www.sigaev.ru/
 


В списке pgsql-hackers по дате отправления:

Предыдущее
От: Teodor Sigaev
Дата:
Сообщение: Re: string_to_array eats too much memory?
Следующее
От: "Heikki Linnakangas"
Дата:
Сообщение: Re: 8.2 Beta 3 Now Available for Download / Testing ...