 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
.NET Forum / Languages / C# / August 2007Regex question
I have a regex question and it never occurred to me to ask here, until I saw Jesse Houwing's quick response to Phil for his Regex question. I have some filenames that I'm trying to parse out of URLs. (href=("|')http://.www\.thesite\.com/.{1,7}/)(?<filename>.[^"|']) This generally works, but the problem is some of the image files have .th.jpg at the end to indicate thumbnails. I want to exclude those. I just want the ones that don't have .th. before the file extension. I've tried using forward and reverse negative lookups, but I guess I'm not using them correctly. Any help on how to get a non-match for a .th. would be great. (href=(?:"|')http://www\.thesite\.com/.+?/)(?<filename>(?:.(?!\.th\.))+?)["|'] A few words of explanation: First, your regular expression has a "." prior to the "www". This indicates that a single non-line-break character *must* precede the "www". I don't know if that's intentional, but I took it out. Second, I'm not sure why you specified any non-line-break character repeated between 1 and 7 times, followed by a slash. That seemed excessive. Regardless of the length, the important thing is that the length is non-zero, and it is followed by a slash. So, I used the dot (non-line-break) with a lazy plus ("+?"). Lazy matching is an important aspect of regular expressions. It means that the match should be repeated as *few* times as possible. The default is "greedy" matching, which indicates that the match should be repeated as *many* times as possible. In a regular expression where a lazy match is followed by some other match, the lazy match stops at the first incidence of the next match. This means that the sequence: .+?/ indicates "any non-line-break character repeated as few times as possible, until a slash is reached, followed by a slash. It wasn't so important in that section of the regular expression, but it's critical to the latter part: (?<filename>(?:.(?!\.th\.))+?)["|'] This is your filename group. You can use quantifiers with un-named groups. So, what the last part says about the "filename" group is: Match any single non-line-break character that is *not* followed by ".th." character sequence as few times as possible. This is followed by ["|'] - which indicates that the next match is either a single quote or a double quote. Therefore, the matching of the "filename" group halts when the quote is reached. In other words, as the group is at the end of the match, it will be followed by a single or double quote, and I use that to indicate the end of that matching group, by using a lazy quantifier.  SignatureHTH, Kevin Spencer Microsoft MVP DSI PrintManager, Miradyne Component Libraries: http://www.miradyne.net >I have a regex question and it never occurred to me to ask here, until I >saw Jesse Houwing's quick response to Phil for his Regex question. [quoted text clipped - 10 lines] > using them correctly. Any help on how to get a non-match for a .th. would > be great. Hello Kevin, Pedrito, I've got a few things to add to what kevin already explained. First, if you're using character sets, there's no need to use a or (|) in there. So either singlequote or doublequote can be written as ["']. Writing ["|'] would actually allow a | on that position as well. Second on the .+? between the two paths. href=["']http://www\.thesite\.com/.+?/)(?<filename>(?:.(?!\.th\.) I'd probably use a more precise expression: /([^/'"]+/)+ This will capture all directories until it can no longer find a trailing slash. It does not use a reluctant (+?) modifier and that usually makes things faster. Also note that I excluded the quotes from the expression. Otherwise there is a chance here it might try to capture the whole HTML file. And last on the exclusing of thumbnails: you're using (.(?!.th.)) to skip the thumbnail extention. It's probably faster and easier to use a negative look behing in this case: ((?!\.th\.)[^'"])+ It will capture any chracter that isn't a singel or double quote. It also makes sure every caracter captured isn't the start of .th.. But here your mileage may vary. There are a few variants which will all accomplish the same thing here: [^'"]+(?<!\.th)\.jpg [^'"]+(?<!.th.jpg) (?!['"]+\.th\.)[^'"]+ You might want to test all possibilities to see which one is faster, or to see which one you find the easiest to read. Also note that negative look behinds aren't supported by all regex libraries (though the standard .NET one will work just fine, Javascript actually won't). Also, if you aren't using this code clientside in Javascript, you might want to see if the System.IO.Path or System.IO.Uri might provide a better way to parse the paths. In that case you can just capture all things that follow href= and do the rest of the vaidation in code (migth be faster, though I'm not sure). Which would lead to the following expression: href=["']http://www\.thesite\.com/([^/'"]+/)+(?<filename>((?!\.th\.)[^'"])+)["'] The last ["'] would even be optional as it's already caught by the filename expression. -- Jesse Houwing jesse.houwing at sogeti.nl > (href=(?:"|')http://www\.thesite\.com/.+?/)(?<filename>(?:.(?!\.th\.)) > +?)["|'] [quoted text clipped - 59 lines] >> I'm not using them correctly. Any help on how to get a non-match for >> a .th. would be great. Hi Jesse, You're of course correct about the character set for the quotes. I overlooked that one in my conversion process. I'm so embarassed! Signature;-), Kevin Spencer Microsoft MVP DSI PrintManager, Miradyne Component Libraries: http://www.miradyne.net > Hello Kevin, Pedrito, > [quoted text clipped - 119 lines] >>> I'm not using them correctly. Any help on how to get a non-match for >>> a .th. would be great. Hello Kevin, > Hi Jesse, > > You're of course correct about the character set for the quotes. I > overlooked that one in my conversion process. I'm so embarassed! No sweat... Happens to the best of us. Jesse > Kevin Spencer > Microsoft MVP [quoted text clipped - 126 lines] >>>> I'm not using them correctly. Any help on how to get a non-match >>>> for a .th. would be great. -- Jesse Houwing jesse.houwing at sogeti.nl Free MagazinesGet these publications absolutely FREE for up to 12 months. There are no hidden fees and no obligation. Simply choose a title, complete the application form and submit it. Read more ...
|
Reply to this Message
|
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2008 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.