Question

I have a string like this:

Please refer to document ABC.123.1234.1234 and document CBA.321.4321

I’m running two different regex searches to separately identify two different document identifiers. The expression for the first identifer works great:

ABC.123.1234.1234 = \b[A-Z]{3}\.\d{1,4}\.\d{1,4}\.\d{1,4}\b

Now, the problem I’m having is with trying to extract the smaller identifier using the following expression:

\b[A-Z]{3}\.\d{1,3}\.\d{1,4}\b

Unfortunately this returns both results, ABC.123.1234 & CBA.321.4321. The only result I require the second expression to return is CBA.321.4321.

Answer 1

Not sure which regex system you are using, since they all have slightly different syntax.

What you want is a negative zero-width lookahead assertion, to make sure you get your match, and matches only count if they aren’t followed by \.[A-Za-z]{4}.

Also, are the numbers in your data actually variable-width? If not, it would be easier to get matches if you match {4} instead of {1,4}. The look-ahead assertion wouldn’t be as easy to implement otherwise.

You could still implement them, though. Simply make your negative look-ahead match \d*\.\d{1,4} (the \d* being the important part to avoid partial matches).

Edit:

Since you’re using VB.Net, here’s the syntax for negative lookahead assertions in that Regex implementation:

(?!subexpression)

So your regex might become something like:

\b[A-Z]{3}\.\d{1,4}\.\d{1,4}(?!\d*\.\d{1,4}\b)

The important part to remove the longer matches, and deal with the variable width numbers is:

(?!\d*\.\d{1,4}\b)