[tpm] New Regex question

Uri Guttman uri at stemsystems.com
Sat Oct 27 15:51:29 PDT 2012


On 10/27/2012 06:19 PM, Chris Jones wrote:
> Thank Stuart, I did end up using $File::Find::dir, very helpful.
>
> New question.  I would like to replace multiple spaces with single
> spaces and multiple end of line  with a single end of line.
>
> I thought this might work:
> $expression =~ s/(\s+)/\1/g; # puts back what it found
>
> So I tried
> $expression =~ s/(\s)/\1\s/g;

someone else gave you a decent solution but i want to clear up a 
misconception you have. the right side of s/// is a replacement string 
and \s is only a regex shortcut for white space. putting \s there will 
only give you a 's' char in the replacement. also \1 is not meant to be 
used in the replacement but only in the regex. $1 is the correct thing 
in the replacement. \1 is supported for backwards compatibility.

another and likely better solution is to use tr///. this code will 
replace any run of space, \t, \r\n with a single one of each.

	tr/ \t\\n\r//s

the replacement pattern will be duplicated from the left side and the /s 
modifier will squeeze them to 1 char. if you want to make any \r or \n 
runs be squeezed to a single \n you need an explicit replacement part:

	tr{ \t\n\r}{ \t\n\n}s

i switched the delims as there were too many /\ leaning toothpicks.

uri



More information about the toronto-pm mailing list