simple regular expression problem

[steveryan]

Hi all,

I have a piece of code which searches a HTML document for movie links using preg_match_all.

Here is the pattern I'm using:

$pattern="([url]http://.+/url|(.mpg)|(.mpeg)|(.wmv)|(.mov))";

For some reason the branching part - (.avi)|(.mpg)|(.mpeg)|(.wmv)|(.mov) - isn't working properly. It will only find AVI's (which is the first branch option). It won't find any of the other movie types.

Do you know why this is? Is here a better way to do branching/optional searching?

Thanks 🙂

[BLottman]

$pattern="([url]http://.+/url|(.mpg)|(.mpeg)|(.wmv)|(.mov))";

Have you tried:

$pattern="([url]http://.+/url|[url]http://.+/url|[url]http://.+/url|[url]http://.+/url|[url]http://.+/url)";

I'm not familiar with branching but it seems you're asking the [url]http://.+[/url] to plow threw each of the separate patterns. Is the pipe an "or" symbol?

[steveryan]

thank you 🙂

[jeremuck]

I believe if you modify the first one so that the movie types are surrounded by [] it will work .

$pattern="([url]http://.+[/url][(.avi)|(.mpg)|(.mpeg)|(.wmv)|(.mov)])";

like that, but I'm not totally sure.

I found an interesting site where people post lots of regexp patterns, FYI.

Try looking for it on google.

[Weedpacket]

jeremuck is almost right; just use () instead of [].

(http://.+(\.(avi|mpg|mpeg|wmv|mov)))

I lobbed one other change in there to make it look neater. More changes could be made (like tacking a $ on the end), but it depends on circumstances.

[jeremuck]

weedpacket's the man.
That looks much better.