Carrying out a full-text search, but without the full-text search... - MySQL

[Daz]

Hi everyone.

I have a fairly simple question, but the answer is likely to be very
complex. I would like to create a wiki that will allow full-text
searches. However, as you know, full-text searches in a large database
is very CPU intensive. I need to look into what options are available
in the way of breaking everything down into keywords, and how I'd do
it efficiently. I know that there is no definitive answer to my query,
but I have no idea where to start looking. I would like to make my
wiki very much like a small-scale search engine. Much like Google, but
far, far smaller by comparison.

Google doesn't just do a full-text search on the several billion pages
it has stored every time you make a query. They evidently have some
algorithm running behind it all which breaks everything down very
efficiently indeed.

I would very much appreciate it if anyone could point me in the right
direction, with any suggestions, whether thy be logic, doents I
should read, or terms I can Google.

Thanks in advance.

Daz.

[Joachim]

Daz schrieb: 

Try a FULLTEXT index.
I have never used it, but it's supposed to do exactly what you want.

Regards,
Jo

[lark]

Joachim Durchholz wrote: 
>
> Try a FULLTEXT index.
> I have never used it, but it's supposed to do exactly what you want.
>
> Regards,
> Jo[/ref]

that's right. with a fulltext index you can do wonders however consider
the limiations:
1-works only in myisam tables
2-collation dependent
3-no column prefixes apply (obviously)

--
lark -- net
To reply to me directly, delete "despam".

[Daz]

On May 1, 8:11 pm, lark <net> wrote: [/ref]
> 
> 
>
> that's right. with a fulltext index you can do wonders however consider
> the limiations:
> 1-works only in myisam tables
> 2-collation dependent
> 3-no column prefixes apply (obviously)
>
> --
> lark -- net
> To reply to me directly, delete "despam".[/ref]

That's excellent. Thanks guys.

I didn't realise you could use a full text index. I assumed that
searching fulltext would be very heavy on resources.

Thanks again.