Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.sai.msu.su/~megera/postgres/gist/tsearch/V2/dicts/README.intdict Дата изменения: Wed Oct 25 23:12:03 2006 Дата индексирования: Sat Dec 22 09:13:34 2007 Кодировка: Поисковые слова: trifid nebula

IntDict - tsearch2 dictionary for integers.

Motivation for this dictionary is to control indexing of integers
(signed and unsigned), and, consequently, to minimize the number of unique
words, which, in turn, greatly affects to performance of searching.

Notes about developing this dictionary using dictionary template
generator (Gendict) are available from
http://www.sai.msu.su/~megera/wiki/Gendict

Dictionary accepts two init options:

* MAXLEN parameter specifies maximum length of the number considered as a 'good' integer.
Default value is 6.
* REJECTLONG parameter specifies if 'long' integer should be indexed or treated as stop word.
o If REJECTLONG=false (default), than dictionary returns prefixed part of integer number with length MAXLEN.
o If REJECTLONG=true, than dictionary consider integer as a stop word.

Examples:

* 12345678902132435454 - 'garbage'
* 123456 - 'good' integer number suitable for indexing

Usage:

1. Compile and install

2. Load dictionary ( assume tsearch2 is already loaded)
psql qq < dict_intdict.sql

3. Test it

qq=# select dict_name, dict_initoption from pg_ts_dict where dict_name='intdict';
dict_name | dict_initoption
-----------+---------------------------
intdict | MAXLEN=6,REJECTLONG=false
qq=# select lexize('intdict','12345678');
lexize
----------
{123456}
qq=# select lexize('intdict','123');
lexize
--------
{123}

4. Specify intdict dictionary to process int and uint tokens
( for simplicity, I did that for all configurations)

qq=# select * from pg_ts_cfgmap where tok_alias ~ 'int';
ts_name | tok_alias | dict_name
-----------------+-----------+-----------
default | int | {simple}
default | uint | {simple}
default_russian | int | {simple}
default_russian | uint | {simple}
simple | int | {simple}
simple | uint | {simple}
qq=# update pg_ts_cfgmap set dict_name='{intdict}' where tok_alias ~ 'int';
UPDATE 6
qq=# select * from pg_ts_cfgmap where tok_alias ~ 'int';
ts_name | tok_alias | dict_name
-----------------+-----------+-----------
default | int | {intdict}
default | uint | {intdict}
default_russian | int | {intdict}
default_russian | uint | {intdict}
simple | int | {intdict}
simple | uint | {intdict}

That's all.