Re: regex for negating a sequence

i think this will work for you:

/\x26\x67(?!\x26\x67)(.(?!\x26\x67)){6}.{2}/s
[download]

Based upon this test code to try to match ab followed by 4 bytes w/no ab sequence:

 perl -le 'print for map { /ab(?!ab)(.(?!ab)){2}.{2}/s ? $& : () } @AR
+GV' abhjabdf abasdasd abdasdadsab ababblah
OUTPUTS:
abasda
abdasd
abblah
[download]

Update: Here's a commented version:

  /
    \x26\x67     # Starting sequence (let's call it AB)
    (?!\x26\x67) # Exclude double starting seqeuence, e.g. ABAB
    (?:          # non-capture
      .(?!\x26\x67) # Any character that's not followed by the sequenc
+e
    ){6}         # Need N-2 of these
    .{2}         # And the last two characters can be anything.
                 # note we know they're not AB because otherwise the
                 # (N-2)th item in the previous part wouldn't have mat
+ched
  /sx            # Enable comments, and make . mean everything
[download]

Update: Here's a generic version:

  my $marker = '&g';
  my $N = 8;
  my $n1 = $N-2;
  /$marker(?!$marker)(?:.(?!$marker)){$n1}.{2}/s
[download]

Update: (As pointed out by diotalevi) Added the /s modifier in case the data stream can have a newline in it.

Comment on Re: regex for negating a sequence Select or Download Code

Replies are listed 'Best First'.
Re^2: regex for negating a sequence by ikegami (Patriarch) on Oct 03, 2006 at 19:35 UTC
Your regexp can be simplified to the following: `/\x26\x67(?:(?!\x26\x67).){7}./s` [download] If you'd rather have simplicity at the cost of an extra check, use the following: `/\x26\x67(?:(?!\x26\x67).){8}/s` [download] `/(?:(?!$re).){8}/` is the equivalent of `/[^$chars]{8}/` but it (negatively) matches entire regexps instead of a choice of characters.	[reply] [d/l] [select]
Re^3: regex for negating a sequence by ysth (Canon) on Oct 03, 2006 at 20:25 UTC
Another alternative: `/\x26\x67(?!.{0,6}\x26\x67).{8}/s` [download] I've interpreted "does not appear" differently than you when the eighth char is the \x26 of a \x26\x67 sequence.	[reply] [d/l]
Re^3: regex for negating a sequence by davidrw (Prior) on Oct 03, 2006 at 20:40 UTC
nice. My first attempt was trying to do it like that, but i started with something like: `/\x26\x67(.(?!\x26\x67)){7}./s` [download] which doesn't work for the case of consectutive sequences.. i just didn't think of putting the look-ahead behind the dot :)	[reply] [d/l]
Re^2: regex for negating a sequence by diotalevi (Canon) on Oct 03, 2006 at 18:36 UTC
Your . is excluding the \n character. Use the /s flag in some form if you really mean "any character." ⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊	[reply]


Problems? Is your data what you think it is?
	PerlMonks