Regular expression to locate string within 3 words

[ADILT]

Need your help DLP and regular expression gurus, I am attemtping to use a data identified to locate US Passport numbers, but only when the  value matched is within 2 to 3 words of the word (Visa or Passport). Below is my attempt, works great in any regular expression tester (http://www.regextester.com/)

 

My expression so far:

\b(?:c[0-9a-zA-Z]{6,9}(?:\W+\w+){1,2}?\W+visa|visa(?:\W+\w+){1,2}?\W+c[0-9a-zA-Z]{6,9})\b

 

Test data (When testing it workds to find the first 4 words as a match):

c123456df at the visa   passport

 

So although this works in the regular expression tester it does not work in DLP data identifier or directly into policy. I am sure DLP does not use Javascript regular expressions like we would expect. I located an artcle which states that it may use Boost, but not sure how to test Boost.

 

Please help

[stephane.fichet]

hello,

 dont know if it is a typo or not i your question..do you try to use a data identifier ?

Take care that data identifier patterns does not use all regexp capabilities. (seems to me there some input on that in DLP admin guide).

Dont know which DLP version you are using, but on 12.5.0, using \b in data identifier pattern make crash our DLP system (message processing crashed

and no policies are applied).

if you dont have any validator or need for uniqueness, you should better use a directly a regexp rule

 

DLP is using perl regular expression, you could find some input in this article :

https://support.symantec.com/en_US/article.HOWTO53607.html

 

you may test something like that in DLP:

c[0-9a-zA-Z]{6,9} (\w+)\s(\w+)((\s(\w+))?) (visa|passport)

 

 regards

[ADILT]

Stephane, using the regex directly rather than via a data identifier, works. Thank you very much.