Complex RegEx in AWK errors on execution

[bdmeyer]

This is working from the command line:
grep -P --color '\b(?!000)(?!666)(?:[0-6]\d{2}|7(?:[0-356]\d|7[012]))[- ](?!00)\d{2}[- ](?!0000)\d{4}\b' *.txt

The --color shows me it is matching exactly.

What I really need though is to not have the entire line returned. I only need the exact matches returned from inside each line. (like what color is showing me)

So I tried:

awk '/\b(?!000)(?!666)(?:[0-6]\d{2}|7(?:[0-356]\d|7[012]))[- ](?!00)\d{2}[- ](?!0000)\d{4}\b/ { print $0 }' file.text

and I get this error:

awk: illegal primary in regular expression[/color red] \b(?!000)(?!666)(?:[0-6]\d{2}|7(?:[0-356]\d|7[012]))[- ](?!00)\d{2}[- ](?!0000)\d{4}\b at[/color red] !000)(?!666)(?:[0-6]\d{2}|7(?:[0-356]\d|7[012]))[- ](?!00)\d{2}[- ](?!0000)\d{4}\b
source line number 1
context is[/color red]
/\b(?!000)(?!666)(?:[0-6]\d{2}|7(?:[0-356]\d|7[012]))[- ](?!00)\d{2}[- >>> ](?!0000)\d{4}\b/ <<< [/color red]


I believe it ha something to do with either the backtics, or more likely the '!' in the regex.

I am not hung up on using awk, I would be happy to use any tool that might parse just the matched values instead of the entire matched line, so I can dump that to a file.

Any Help, is appreciated:
--Bruce D. Meyer

 

[feherke]

Hi

[tt]\b[/tt] in [tt]awk[/tt] is a backspace character.

In [tt]gawk[/tt] the start of the word is [tt]\<[/tt] . As far as I know, other [tt]awk[/tt] implementations can not match word boundaries.

But while you are using [tt]grep[/tt] with the -P ( --perl-regexp ) switch, would be better to move to [tt]perl[/tt] instead. Then your regular expression should work :

perl -nae 'print $F[0] if /\b(?!000)(?!666)(?:[0-6]\d{2}|7(?:[0-356]\d|7[012]))[- ](?!00)\d{2}[- ](?!0000)\d{4}\b/' file.text

Sorry for the off-topic suggestion.

Feherke.

http://free.rootshell.be/~feherke/