java - break url into parts

If the url is malformed, Why not just ignore the exception and proceed to other url's,

You can make a method where you cvatch Exception and ignore it.

Something like that, though you might consdier
differentiate if it has some other Exception not URLMalformed
and do somthing about it , or not:

public URL myDecode(String url){
try{
URL u = new URL(URLDecoder.decode(url, "UTF-8"));
} catch(Exception ex) {
return null;
}
return u;
}

That'll be like you asked for - null instaed of exception

I need to prevent the exceptions from being formed. I can wrap them in try / catch.

The act of creating and throwing an exception is a performance issue.

You can look and amnalyze most of your problesm - hwat is their cause
and then do preliminary check with substrings

I'm also interested in a way of doing this without creating a new URL object for each url.

>>You can look and amnalyze most of your problesm - hwat is their cause
and then do preliminary check with substrings

There are a lot of cases. I don't want to check them by hand. I'm hoping for a premade something that already does this.

New odd cases will continuously spring up; it needs a solution more robust than the hackery I'd turn out in a few hours.

Well, you can also find the
source code for URLDecoder and modify it so it does not create
Exception

There iit is
http://www.java2s.com/Open-Source/Java-Document/6.0-JDK-Modules/j2me/java/net/URLDecoder.java.htm

Can't believe it is the actual source code - so short.
But maybe can still be useful to you

Url is a much bigger piece of the puzzle. I need to extract protocol domain path and querystring separately while identifying bad urls.

Url is a much bigger piece of the puzzle. I need to extract protocol domain path and querystring separately while identifying bad urls.

Try the following regular expression.  I don't know how efficient this will be -- might be faster to use the URL class and handle the exceptions, but you can performance test it.  You should be able to match any valid URL except mailto: and ftp://username:password@ with this expression and, once successfully matched, you can use the Matcher.group() method to retrieve the URL parts or null/empty string if they did not exist (see the javadoc for Matcher.group(), it will return an empty string if the group could match the empty string.  I just put this together and haven't tested it so there may be some errors, but hopefully its a starting point.

If you are parsing private URLs with a non-standard TLD, you can replace the TLD section with the expression below with: (?:[a-zA-Z][a-zA-Z0-9\\-]*[a-zA-Z0-9])

(?:([a-zA-Z]*)://)((?:(?:[0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|25[0-5])\\.){3}(?:[0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|25[0-5]))|(?:[a-zA-Z0-9\\-]*[a-zA-Z0-9]\\.)+(?i:aero|asia|biz|com|coop|edu|gov|info|int|jobs|mil|mobi|museum|name|net|org|pro|tel|travel|xxx|[a-z]{2}))(?::(0?\\d{1,4}|6[0-4]\\d{3}|65[0-4]\\d{2}|655[0-2]\\d|6553[0-5]))?(?:((?/[^#?])*)(?:#([a-zA-Z0-9]*))?(?:\\?((?:[a-zA-Z]\\w*=[^&]*&)*(?:[a-zA-Z]\\w*=[^&]*))))?

Select allOpen in new window

Quick explanation, with group indexes.

Select allOpen in new window

(?:([a-zA-Z]*)://) # matches protocol, if present, in format "xyz://", GROUP 1
(
  (?:
    (?:[0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|25[0-5])\\.){3}
    (?:[0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|25[0-5])
  ) # match IPv4 address
  | # OR DNS name
  (?:[a-zA-Z0-9\\-]*[a-zA-Z0-9]\\.)+
  (?i:aero|asia|biz|com|coop|edu|gov|info|int|jobs|mil|mobi|museum|name|net|org|pro|tel|travel|xxx|[a-z]{2})
) # server address, GROUP 2
(?::(0?\\d{1,4}|6[0-4]\\d{3}|65[0-4]\\d{2}|655[0-2]\\d|6553[0-5]))? # matches optional port 0-65535, GROUP 3
(?:
  ((?/[^#?])*) # matches path, if present, GROUP 4
  (?:#([a-zA-Z0-9]*))? # matches hash, if present, GROUP 5
  (?:\\?((?:[a-zA-Z]\\w*=[^&]*&)*(?:[a-zA-Z]\\w*=[^&]*))) # matches query string, if present, GROUP 6
)?

Select allOpen in new window