Doing Searches On Articles?

Dark_Phoenix

Okay, I'm trying to find the best way of doing a search on a group of articles/posts/whatever (basically, anything with a lot of text).

Now, the idea I had to do it was to split the file on " " (space), weed out average words like the, a, an, etc., then save them in a LONGTEXT in the database like this: test:5,fun:10,organ:2, you get the idea. Then when the user searches, I just need to use a LIKE statement, get the number after the : and the id of the article, then sort them based on the number.

Is this an efficient way of searching, or is there a faster/better/whatever way?

jayant

store the complete this in mysql.
use full text search feature of mysql (see mysql.com docs) to query.
mysql will auto ignore words having less than 4 letters/strip common words. and will sort the results for you in order of importance also. 🙂

Dark_Phoenix

Is that the fastest way of doing fulltext searches in a pure PHP system?

jayant

you have 3 options

1) convert all data into mysql and use mysql's powerfull full text search.

2) use some readymade search script, look hotscripts.com/PHP/

3) make one urself. we will be happy to help you build it, but dont expect us to build it for u.

Dark_Phoenix

I'm not asking anyone to build anything. I'm asking for the fastest method of doing fulltext searches on articles using pure PHP.

jayant

fastest will be mysql method
why?
heres why:
mysql is a dbms, and it needs quick and fast methods for performance reasons. and lot of optimizations etc are already done for the same.

seby

you really cant get much faster than mysql. its really fast in searches.

i have a database with 5 million entries and i can perform a full search (a complete script in php) in less than 0.50 seconds...

and it doesn't kill the server at all.. if you did a filesystem search with 5 million lines you'll probably make the server unreachable for 15 minutes 😉

Dark_Phoenix

Originally posted by seby
you really cant get much faster than mysql. its really fast in searches.

i have a database with 5 million entries and i can perform a full search (a complete script in php) in less than 0.50 seconds...

and it doesn't kill the server at all.. if you did a filesystem search with 5 million lines you'll probably make the server unreachable for 15 minutes 😉

Okay. I was just reading at mysql.com that fulltext searching wasn't very fast, so I wanted to make sure 😉