Re: counting overlapping patterns

The key is to match without consuming text. The following works.

while("AAAA" =~ /(?=AA)/g){
  $count++;
}
[download]

It'll go through it repeatedly, starting at each position, but never consuming anything.

Comment on Re: counting overlapping patterns Download Code

Replies are listed 'Best First'.
Re^2: counting overlapping patterns by ikegami (Patriarch) on Feb 18, 2005 at 21:46 UTC
Alternatively, `while ('AAAA' =~ /A(?=A)/g) { $count++; }` [download] which looks slightly less weird to me.	[reply] [d/l]
Re^3: counting overlapping patterns by Eimi Metamorphoumai (Deacon) on Feb 18, 2005 at 22:15 UTC
But requires rewriting the pattern. Mine will work even if the pattern comes from a variable, or contains different possibilities for the first character ('`/AA\|BB/`' becomes '`/(?=AA\|BB)/`'). My approach was to try to do minimal tinkering with the original, but yeah, it's not as obvious what's actually going on.	[reply] [d/l] [select]
Re^4: counting overlapping patterns by ikegami (Patriarch) on Feb 18, 2005 at 23:39 UTC
Good point.	[reply]


Your skill will accomplish what the force of many cannot
	PerlMonks