how to rebuild phpbb search engine

zlisiecki · Post by **zlisiecki** » Fri Jan 08, 2016 6:27 pm

Hi, my search tables are quite big after 10 years forum activity.

phpbb_search_wordlist 394,235 MyISAM utf8_bin 27.2 MB with 20 MB for index !
phpbb_search_wordmatch 5,336,358 MyISAM utf8_bin 304.1 MB

My phorum is devoted to certain disease and i assume some 4 k terms is far enough. The rest (nearly 400 k now) regards the human language generally and not the disease only. The wordlist is obviously too big, the search eats too much processor power and it cannot be used any more.

Is there any possiblity to clean both tables, rescan the whole forum and build them again ?

Another question regards how to limit search engine generally only to some subfora. Does anybody know how to acomplish this ? I suppose writing an extenssion won't be too difficult.

Post by **Brf** » Fri Jan 08, 2016 6:44 pm

In Manage Forums you can turn off Search Indexing in the settings for individual forums.
You can rebuild your search index by going into Maintenance / Search Index and delete the old index and re-create it. I do not know of a way you could select which words to index.

zlisiecki · Post by **zlisiecki** » Fri Jan 08, 2016 8:17 pm

Thank you Brf, .I'v overseen this possibility. I have set popularity treshold in the server configuration 10 % . Still I'll have nearly 100 000 words. The option could make sense to delete some words like "thank you".

Post by **Brf** » Fri Jan 08, 2016 8:23 pm

If you want to reduce the number of words, you would reduce the popularity threshold, not increase it.

zlisiecki · Post by **zlisiecki** » Fri Jan 08, 2016 8:58 pm

Oh, I'v misundestood this. I thought for 0 % treshhold all words are included and the higher the treshhold the more popular words fall into this limit and won't be included. Is't it so ?

Post by **Brf** » Fri Jan 08, 2016 9:02 pm

By default, any word that appears in more than 5% of the posts is excluded.
You changed it to 10%, which means the word would have to appear in 10% of the posts, which is twice as many posts. You want to reduce it to 1%. Then it would exclude any words which appear in more than 1% of the posts.