free open-source SQL full-text search engine

Need a hand?
+1-888-333-1345


Sphinx

Community

Services

Misc

 Subscribe in a reader

Tracked by ClickAider

Forums :: Register :: Login :: Forgot your password? :: Search

anonymous user


omg why sphinx making so big search mistakes

Common forum | 1 | 2 | 3 | 4 | 5 | ... | 263 | 264 | 265 | 266 | next »» | Create new thread

sraken666

Name: sraken
Posts: 41

2010-01-04 13:10:49 | reply!


searching phrase: 321755276
result found in this text:
[code:1:c77c043ee7]http://rapidshare.com/files/90009145/Marie-Antoinette_2006_DvDrip_E
ng_-aXXo.part1.rar
http://rapidshare.com/files/90020660/Marie-Antoinette_2006_DvDrip_Eng_-aXXo.part2.rar
http://rapidshare.com/files/90030636/Marie-Antoinette_2006_DvDrip_Eng_-aXXo.part3.rar
http://rapidshare.com/files/90039250/Marie-Antoinette_2006_DvDrip_Eng_-aXXo.part4.rar
http://rapidshare.com/files/90054035/Marie-Antoinette_2006_DvDrip_Eng_-aXXo.part5.rar
http://rapidshare.com/files/90062944/Marie-Antoinette_2006_DvDrip_Eng_-aXXo.part6.rar
http://rapidshare.com/files/90070683/Marie-Antoinette_2006_DvDrip_Eng_-aXXo.part7.rar
http://rapidshare.com/files/90072554/Marie-Antoinette_2006_DvDrip_Eng_-aXXo.part8.rar[
/code:1:c77c043ee7]

WHAT THE HELL? Help, so many sphinx search mistakes.

barryhunter

Name: Barry Hunter
Posts: 1443

to: sraken666, 2010-01-04 23:19:20 | reply!


We don't know your sphinx config, but I would think there is mistake somewhere. Sphinx
itself is reliable for most people.

sraken666

Name: sraken
Posts: 41

to: barryhunter, 2010-01-05 10:15:25 | reply!


> We don't know your sphinx config, but I would think there is mistake somewhere. Sphinx
> itself is reliable for most people.

This is config of this index where this false match appears:

index phpbb_posts_text
{
                source = main
                path = /usr/mysql/data/phpbb_posts_text
                docinfo = extern
                mlock = 0
                morphology = none
                min_word_len = 1
                charset_type = utf-8
                html_strip = 0
}

source main
{
                type = mysql
                sql_host = localhost
                sql_user = root
                sql_pass = password
                sql_db = database
                sql_port = 3306
                sql_ranged_throttle = 0


        sql_query_pre = REPLACE INTO phpbb_sph_counter SELECT 1, MAX(post_id) FROM
        phpbb_posts_text
        sql_query = SELECT pt.post_id, pt.post_id AS pid, pt.post_subject, pt.post_text FROM
        phpbb_posts_text pt\
                WHERE pt.post_id<= ( SELECT max_doc_id FROM phpbb_sph_counter WHERE counter_id=1 )
                sql_attr_uint = pid
}

indexer
{
                mem_limit = 512M
                  max_iops = 40
                  max_iosize = 1048576
}

searchd
{
                port = 3312
                log = /var/log/searchd.log
                query_log = /var/log/query.log
                read_timeout = 5
                max_children = 30
                pid_file = /var/log/searchd.pid
                max_matches = 5000
                seamless_rotate = 1
                preopen_indexes = 0
                unlink_old = 1
}

Here it is, something here which could cause this false matches?

sraken666

Name: sraken
Posts: 41

to: sraken666, 2010-01-09 16:14:53 | reply!


refresh

dochead

Name: Shayan Raghavjee
Posts: 30

to: sraken666, 2010-01-11 10:18:35 | reply!


Some code would be more useful than a bump.

You could be using the wrong search mode ferinstance.

sraken666

Name: sraken
Posts: 41

to: dochead, 2010-01-14 23:00:03 | reply!


> Some code would be more useful than a bump.
>
> You could be using the wrong search mode ferinstance.
>

Tryed all search modes, in every same mistake.

sraken666

Name: sraken
Posts: 41

to: sraken666, 2010-02-09 12:01:59 | reply!


refresh of problem new example

[root@server]# /usr/local/bin/search -i posts_text -p 321755276
Sphinx 0.9.9-release (r2117)
Copyright (c) 2001-2009, Andrew Aksyonoff

using config file '/usr/local/etc/sphinx.conf'...
index 'posts_text': query '321755276 ': returned 3 matches of 3 total in 0.014 sec

displaying matches:
1. document=2018230, weight=1, pid=2018230 [ I N C O R R E C T !!! ]
2. document=10934921, weight=1, pid=10934921 [ C O R R E C T ]
3. document=10935621, weight=1, pid=10935621 [ C O R R E C T ]

words:
1. '321755276': 3 documents, 4 hits

In posts_id's 10934921 and 10935621 there is in indeed phrase inside 321755276 but on
first match post_id 2018230 there is not search prahse 321755276 but still displayed as
search match.

This error are quite often when searching by various id's (not words).

sraken666

Name: sraken
Posts: 41

to: sraken666, 2010-02-15 08:22:17 | reply!


problem still exist and its serious, quite many false matches appearing

barryhunter

Name: Barry Hunter
Posts: 1443

to: sraken666, 2010-02-15 11:10:02 | reply!


Are you reindexing propelly? Maybe you using a old index.

sraken666

Name: sraken
Posts: 41

to: barryhunter, 2010-02-15 19:27:42 | reply!


> Are you reindexing propelly? Maybe you using a old index.

nope, im reindexing every day main index, then deltas is gathered all day, then again
main indexing from scratch at night created and rest of day deltas. (like in sphinx guide)

sraken666

Name: sraken
Posts: 41

to: sraken666, 2010-02-18 08:56:02 | reply!


Really noone has any ideas how to TRY to fix it? Me neither :P.
I mean i could edit this post text in DB..... try creating whole diffirent new_index with
similiar data and check if that mistake is there also.
Diffirent indexing methods? Maybe some characters are treated like they should not and
that causing problems.

Advices welcomed.

sraken666

Name: sraken
Posts: 41

to: sraken666, 2010-02-22 18:16:14 | reply!


problem still exist, help

barryhunter

Name: Barry Hunter
Posts: 1443

to: sraken666, 2010-02-22 18:52:24 | reply!


Well I dont know how you expect anyone to help you. We would need more details.

http://catb.org/~esr/faqs/smart-questions.html

sraken666

Name: sraken
Posts: 41

to: barryhunter, 2010-02-23 03:10:21 | reply!


> Well I dont know how you expect anyone to help you. We would need more details.
>
> http://catb.org/~esr/faqs/smart-questions.html

Then tell me, what details do you need?

sraken666

Name: sraken
Posts: 41

to: sraken666, 2010-03-30 13:29:23 | reply!


refresh

Common forum | 1 | 2 | 3 | 4 | 5 | ... | 263 | 264 | 265 | 266 | next »» | Create new thread


Copyright © Sphinx Technologies Inc, 2009