Regex help - URL validation

Hi All,

Thanks in advance.

[((http)|(https))://][www.]domainname.((com)|(gov))[/dir or page name]

I am looking for a regex for the above URL format. Could anyone help me please?

Thanks!
Jenni

Find more posts tagged with

Comments

ISCBorisB

Hi All,

Thanks in advance.

[((http)|(https))://][www.]domainname.((com)|(gov))[/dir or page name]

I am looking for a regex for the above URL format. Could anyone help me please?

Thanks!
Jenni

A formal answer to your request is something like ^(https?:\/\/)?(www\.)?.+?\.(com|gov)(\/.*)?
Please note however that "httkaka:/foo.gov" will match perfectly, after all who's to say that "httkaka:/foo"
is not a valid "domainname" in your terminology?

So, what are you trying to accomplish? Note that formal URL Syntax validation is surprisingly hard to do in a single
regex. If that's what you are trying to do perhaps little script may serve you better

nipper

Bo's comments are spot on. I was looking at this before & it is ugly at best. Of course when I am, we use friendly URLs, specific pages and then anchors within a page. So that got incredibly ugly.

Andy

Jenni

A formal answer to your request is something like ^(https?:\/\/)?(www\.)?.+?\.(com|gov)(\/.*)?
Please note however that "httkaka:/foo.gov" will match perfectly, after all who's to say that "httkaka:/foo"
is not a valid "domainname" in your terminology?

So, what are you trying to accomplish? Note that formal URL Syntax validation is surprisingly hard to do in a single
regex. If that's what you are trying to do perhaps little script may serve you better

Yes Boris, this is what we want. But it is also accepting the following

http://www.domain.comabc

Is there a way I could restict to .com or .gov?

Thanks!
Jenni

nipper

maybe ^(https?:\/\/)?(www\.)?.+?\.(com\/|gov\/)(\/.*)?

ISCBorisB

maybe ^(https?:\/\/)?(www\.)?.+?\.(com\/|gov\/)(\/.*)?

That will work but it is a bit too restrictive. It'll require slash at the end even if there is no path. Following may be somewhat better:

^(https?:\/\/)?(www\.)?.+?\.(com|gov)($|\/.*)

Adam Stoller

That will work but it is a bit too restrictive. It'll require slash at the end even if there is no path. Following may be somewhat better:

^(https?:\/\/)?(www\.)?.+?\.(com|gov)($|\/.*)

Personally - I prefer to use some character other than '/' as a delimiter for my regex's when I have '/' within them - how about:

... =~ m{^(https?://)?[^/]+\.(com|gov)(|/.*)$};

or:

... =~ m{^(?:https?://)?[^/]+\.(?:com|gov)(|/.*)$};

For those interested in some of the fancier aspects of the above:

If the intent was to ensure that "domainname" was part of the string and that it was either "domainname.****" or "www.domainname.****" then it should probably be one of:... =~ m{^(https?://)?(www\.)?$domainname\.(com|gov)(|/.*)$};... =~ m{^(?:https?://)?(?:www\.)?$domainname\.(?:com|gov)(|/.*)$};

ISCBorisB

[QUOTE=ghoti]Personally - I prefer to use some character other than '/' as a delimiter for my regex's when I have '/' within them - how about:

... =~ m{^(https?://)?[^/]+\.(com|gov)(|/.*)$};

or:

... =~ m{^(?:https?://)?[^/]+\.(?:com|gov)(|/.*)$};

For those interested in some of the fancier aspects of the above:

Fish, you are making an assumption that this regex will be used in the Perl context. If that's the case all is fine, but we do not know that. If that is some TS configuration File or DCT Validation regex it is much safer to be conservative, use escaped forward slash, not use "Suppress backreference" etc... And do not eventry to use lookarounds. TS does all that in Java and what Pattern Class they are using is one of their silly little secrets ( looks like NOT Perl5Pattern )

Jenni

Hi All,

Thanks a lot for all the replies.

Here is the code, which we feel was more apt to our req. for the URL validation.

validation-regex="^(((http)|(https))(:\/\/))?(www\.)?[a-zA-Z0-9][\w\.-]*[a-zA-Z0-9]+\.((com)|(gov))(((\/)|(\\))($|[a-zA-Z0-9\.\\\+]+))*$"/>/>

It works for the following,

http://www.yahoo.com
http://www.yahoo.com/
https://www.yahoo.com
www.yahoo.com
http://yahoo.com
http://www.yahoo.com/a/a.htm
http://www.yahoo.com/a.htm

Goti,

Thanks for the idea. Your regex allows the URL of type,
httpsss://www.yahoo.com/a.htm
which is not a valid URL

Thanks!
Jenni

ISCBorisB

Here is the code, which we feel was more apt to our req. for the URL validation.

validation-regex="^(((http)|(https))(:\/\/))?(www\.)?[a-zA-Z0-9][\w\.-]*[a-zA-Z0-9]+\.((com)|(gov))(((\/)|(\\))($|[a-zA-Z0-9\.\\\+]+))*$"/>/>

Is there a question somewhere here?

If not, please note that your regex will "fail" some syntactically correct entries like
www.yahoo.com/hi-jenny and accept some syntactically wrong ones: www.yahoo.com/hi+jenny.

Have you thought about a port that differs from 80? ftp/file/mail protocols? CGI Parameters? Document Anchors?
Permutations of all that? I once did a regex few hundred symbols long for that - it still had
plenty of holes in it

Jenni

Is there a question somewhere here?

If not, please note that your regex will "fail" some syntactically correct entries like
www.yahoo.com/hi-jenny and accept some syntactically wrong ones: www.yahoo.com/hi+jenny.

Have you thought about a port that differs from 80? ftp/file/mail protocols? CGI Parameters? Document Anchors?
Permutations of all that? I once did a regex few hundred symbols long for that - it still had
plenty of holes in it

There is no question. I got the req. regex. Yeh there are some loop holes....

Thanks!
Jenni

Adam Stoller

...
Ghoti,

Thanks for the idea. Your regex allows the URL of type,
httpsss://www.yahoo.com/a.htm
which is not a valid URL

Thanks!
Jenni

If it lets that through it's a symptom of the regex engine being used not understanding that 'x?' means 0-or-1 'x' -- which goes to prove ISCBorisB's point that I was presuming perl code and not a validation-regex directive.