Search patern within pattern in a single substitute command

shobhit has asked for the wisdom of the Perl Monks concerning the following question:

Hi,

I have an input file with 5 columns separated by single space. The first column is of width n (4 in this case). The second column is of width m (4 in this case). The rest of the columns are of variable width. I need to change the ith character (2 in this case) in second column (i<m) make it 'Y' say AND APPEND it as a third column.

Input file

col1 col2 col3 col4 col5
1234 5678 9012 3456 7890
[download]

Output file

col1 col2 cYl2 col3 col4 col5
1234 5678 5Y78 9012 3456 7890
[download]

The following program does it just fine.

use strict;
use warnings;
my $line;
while($line=<>){
        chomp $line;
        $line =~ s/(.{5})(.{4})(.*)/$1$2 $2$3/;
        $line =~ s/(.{11}).(.*)/$1Y$2/;
        print $line."\n";
}
[download]

My question is, is it possible to accomplish this in one s/// command? We can use \1 and \2 etc to reference the pattern we just matched, but I can't figure out how to do it here. Is it even possible?

Thanks for any help with this.

Comment on Search patern within pattern in a single substitute command Select or Download Code

Replies are listed 'Best First'.
Re: Search patern within pattern in a single substitute command by BrowserUk (Patriarch) on Mar 04, 2011 at 01:06 UTC
Here's one way: `#! perl -sw use strict; s[(?:\S+ ){2}\K((.).(..))][$2Y$3 $1], print while <DATA>; =output C:\test>junk56 col1 col2 cYl3 col3 col4 col5 1234 5678 9Y12 9012 3456 7890 =cut __DATA__ col1 col2 col3 col4 col5 1234 5678 9012 3456 7890` [download] Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error. "Science is about questioning the status quo. Questioning authority". In the absence of evidence, opinion is indistinguishable from prejudice.	[reply] [d/l]
Re^2: Search pattern within pattern in a single substitute command by toolic (Bishop) on Mar 04, 2011 at 01:37 UTC
Very cool. Your output doesn't quite match the OP's, but with this slight adjustment, we do: `use warnings; use strict; s[.{5}\K((.).(..))][$1 $2Y$3], print while <DATA>; __DATA__ col1 col2 col3 col4 col5 1234 5678 9012 3456 7890` [download] prints: `col1 col2 cYl2 col3 col4 col5 1234 5678 5Y78 9012 3456 7890` [download]	[reply] [d/l] [select]
Re^2: Search patern within pattern in a single substitute command by shobhit (Sexton) on Mar 04, 2011 at 04:57 UTC
Looking back now, I knew nested pattern `((.).(..))` and should have thought about it here. It turned out to be the key. What does the \K mean? If I don't include it I get o/p `col2 cYl2 col3 col4 col5 5678 5Y78 9012 3456 7890` [download] W/o `\K` the following works too, but I would like to understand \K `s[(.{5})((.).(..))(.*)][$1$2 $3Y$4$5], print while <DATA>;` [download] This is something I would have written if I had thought about the nested pattern in the first place....	[reply] [d/l] [select]
Re^3: Search patern within pattern in a single substitute command by BrowserUk (Patriarch) on Mar 04, 2011 at 05:24 UTC
What does the \K mean? from Perldoc perlre: "Character Classes and other Special Escapes" `\K Keep the stuff left of the \K, don't include it in $&` [download] It says that the bit of the regex preceding it is only used to locate the part of the string to be replaced, but isn't a part of the substitution. It just saves the need for an extra capture. Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error. "Science is about questioning the status quo. Questioning authority". In the absence of evidence, opinion is indistinguishable from prejudice.	[reply] [d/l]
Re^4: Search patern within pattern in a single substitute command by shobhit (Sexton) on Mar 04, 2011 at 05:40 UTC
Re: Search pattern within pattern in a single substitute command by toolic (Bishop) on Mar 04, 2011 at 01:02 UTC
is it possible to accomplish this in one s/// command? Yes, but it is hideously ugly, and you shouldn't do it this way (update... do it BrowserUk's way): `use warnings; use strict; while (my $line = <DATA>){ chomp $line; $line =~ s/(.{5})(.{4})/$1 . $2 . ' ' . substr($2, 0, 1) . 'Y' . s +ubstr($2, 2, 2)/e; print $line . "\n"; } __DATA__ col1 col2 col3 col4 col5 1234 5678 9012 3456 7890` [download] prints: `col1 col2 cYl2 col3 col4 col5 1234 5678 5Y78 9012 3456 7890` [download]	[reply] [d/l] [select]