Counting the number of replaces

[rsriram]

Hi, I am replacing a string throughout a markup file

$_ =~ s/<I>/<italic>/g;

Is there a function/command by which I can count the number of replaces made?

[McDarren]

Yes, just evaluate your regexp in scalar context, eg:

$num_replacements = $_ =~ s/<I>/<italic>/g;

Cheers,
Darren :)

[GrandFather]

Note too that the $_ =~ is redundant. The regex on its own is quite sufficient: my $count = s/<I>/<italic>/g;.

[pingo]

In the spirit of TIMTOWTDI:

s/<I>/$count++;"<italic>"/eg;
print $count || 0;

No, I didn't say it was pretty. ;-)

[cognizant]

Try this. $a = $b=~s/<i>/<italic>/g; print "$a"; Gives the output you need.

[monkey_boy]

I know this is only a quick example, but in general i would avoid using the special variables $a and $b

[Moron]

If you want to get the number of replacements made in the whole file, not just for a particular line, then these should accumulate e.g. using += :

while( <$inputfh> ) {
    $replcount += s/replacethis/withthis/g;
    print $outputfh $_;
}

[shmem]

rsriram, please read How do I post a question effectively?. Your could have found the answer for your last questions by yourself, skimming through perfunc (perl function reference) and perlop (perl operator reference).

cheers,
--shmem

[ForgotPasswordAgain]

I noticed that I had an incorrect superstition about using <I>, namely the brackets, in the left side of s///, as I thought it was normally replaced by reading from a filehandle. It seems to work here, though, so I wonder why I believed that....

[dsheroh]

Because Perl is rather context-sensitive and uses <I> to mean that in other contexts? In the code:

while (<I>) {
  s/<I>/<italic>/g;
}

the first <I> reads from a filehandle and the second is literal text to replace. Fun, isn't it? :)