Question

I have inherited a perl script that pulls data out of some files. The whole script works fine but recently some engineers have been putting in more than one number for a certain spot that usually took one number, so the output is not showing all of what is expected.

Sample input:

CRXXXX: "Then some text"

CRs XXXX, XXXX, XX, XXX

CRXXX "Some Text"

Currently this regex statement I have pulls out the number after the CR, but if given then second line of sample input it prints "s XXXX, XXXX, XX, XXX" instead of the wanted "XXXX XXXX XX XXX"

I am very new to perl and am struggling to figure out how to alter this regex to work on all of the inputs.

 $temp_comment =~ s/\s[cC][rR][-\s:;]*([\d])/\n$1/mg;

Thanks in advance!

Brock

Answer 1

For sample data like:

my $temp_comment =
'CR1234: "Then some text"
 CRs 2345, 3456, 45, 567
 CR678 "Some Text"';

try:

$temp_comment =~ s/(,)|[^\d\n]+/$1?' ':''/semg;

or, if you want to stay close to the string templates:

$temp_comment =~ s/ ^                 # multi-line mode, line start
                    \s*               # leading blanks?
                    CR                # CR tag
                    \D*               # non-number stuff
                     (                  # start capture group
                      (?:\d+ [,\s]*)+   # find (number, comma, space) groups
                     )                  # end capture group
                    \D*               # skip remaining non-number stuff
                    $                 # multi-line mode, line end
                  /$1/mxg;            # set multi-line mode + regex comments "x" 

but you’d have to remove the commas in the number group in a subsequent step.

$temp_comment =~ tr/,//d;             # remove commas in the whole string

or

$temp_comment =~ s/(?<=\d),(?=\s\d)//g;  # remove commas between numbers '11, 22'

For “single step”, you have to use the /e modifier:

$temp_comment =~ s{ ^                 # line start
                    \s*               # leading blanks?
                    CR                # CR tag
                    \D*               # non-number stuff
                    ((?:\d+ [,\s]*)+) # single or group of numbers
                    \D*               # non number stuff
                    $                 # line end
                  }
                  {do{(local$_=$1)=~y/,//d;$_}}mxeg;

This will, on the data above, result in:

1234
2345 3456 45 567
678 

But really, please use, if possible, the simpler two step approach. The latter regex might be a maintainance nightmare for your successors.