[RESOLVED] Perl style regular expressions help

[dmacman1962]

Hi All,

I am installing Sphider search for our site and I want it to not include any URL with "?osCsid=" in it (to exclude all OsCommerce Session variables) and I am terrible at regex coding. Based on their install text, they show that '*/[a]+/' denotes a string with one or more a's in it.

I thought that */[?osCsid=]+/ was the correct way to have it exclude the oscommerce Session url but it is still indexing them. I posted this question on the Sphider forum, but it looks like their traffic is a lot less than here, so I was hoping to get a pointer to a good regex tutorial. I think preg_match has replaced regex in PHP, but Sphider still uses the perl style expressions so I am stuck using them.

Can anyone tell me if my version is correct?

Also, I am not locked in to using Sphider for our search engine and we are willing to pay (or buy) some other search engine code if you have any suggestions. We have Google Custom Search Engine now, but it does not give us the control we need, so we are replacing it.

Thanks,
Don

[laserlight]

dmacman1962 wrote:

I think preg_match has replaced regex in PHP, but Sphider still uses the perl style expressions so I am stuck using them.

preg_match uses PCRE, which stands for "Perl Compatible Regular Expressions". So, what exactly do you mean by "Perl style"?

dmacman1962 wrote:

I thought that /[?osCsid=]+/ was the correct way to have it exclude the oscommerce Session url but it is still indexing them

It is more likely to be:

*/\?osCsid=.*/

since you want to match "?osCsid=" followed by zero or more characters. I presume the leading * is intentional.

[dmacman1962]

Thanks laserlight. I only included the * at the beginning of the expression. So, I will remove it from your answer. And thanks for explaining preg_match as well.

I will use /\?osCsid=.*/ and like you said, there will be more characters after the = for all the session data.

Don