tags:

views:

178

answers:

5

I thought this would be a simple google search but apparently not. What is a regex I can use in C# to parse out a URL including any query string from a larger text? I have spent lots of time and found lots of examples of ones that don't include the query string. And I can't use System.URI, because that assumes you already have the URL... I need to find it in surrounding text.

A: 

Check out this guy's QueryString builder class -

http://weblogs.asp.net/bradvincent/archive/2008/10/27/helper-class-querystring-builder-chainable.aspx

Microsoft also has a UriBuilder that might help you -

http://msdn.microsoft.com/en-us/library/system.uribuilder.query.aspx

Kris Krause
These look fine for building query strings, but JoelFan wants to identify URLs, not build them.
thetaiko
A: 

RegexLib has lots of useful stuff...

http://regexlib.com/Search.aspx?k=querystring

F3
+1  A: 

This should get just about anything (feel free to add additional protocols):

@"(https?|ftp|file)\://[A-Za-z0-9\.\-]+(/[A-Za-z0-9\?\&\=;\+!'\(\)\*\-\._~%]*)*"

The real difficulty is finding the end. As is, this pattern relies on finding an invalid character. That would be anything other than letters, numbers, hyphen or period before the end of the domain name, or anything other than those plus forward slash (/), question mark (?), ampersand (&), equals sign (=), semicolon (;), plus sign (+), exclamation point (!), apostrophe/single quote ('), open/close parentheses, asterisk (*), underscore (_), tilde (~), or percent sign (%) after the domain name.

Note that this would allow invalid URLs like

http://../

And it would pick up stuff after a URL, such as in this string:

Maybe you should try http://www.google.com.

Where "http://www.google.com." (with the trailing period) would be matched.

It would also miss URLs that didn't begin with a protocol specification (specifically, the protocols within the first set of parentheses. For instance, it would miss the URL in this string:

Maybe you should try www.google.com.

It's very difficult to get every case without some better-defined boundaries.

P Daddy
A: 

Use the ABNF at the end of RFC3986 as a starting point to get it right.

This uses them for URI validation in Python; not what you're looking for, but it should give an idea of the direction you should go in: http://gist.github.com/138549

Mark Nottingham
A: 

Sorry I'm not yet able to add comments, but would like to point out that P Daddy's answer requires a little tweaking:

@"(https?|ftp|file)\://[a-zA-Z0-9\.\-]+(/[a-zA-Z0-9\?\&\=;\+!'\(\)\*\-\._~%]*)*"
Starfield
I can't find anything different except that you reversed the order of upper- and lower-case characters (a no-op), and in so doing, fixed a typo I had where I had `a-Z` (lower-case 'a' to upper-case `Z`). Next time, it would be simpler to just point out the typo. I'll fix it.
P Daddy