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

As a follow up to a previous question - I am (now) writing a perl script to do a search and replace (the one liners became too cumbersom) and the issue that is killing me is where the string I am searching for (in order to replace it) the string is split over two lines ... bit like this</P. 

hello there this is a bad <?string?> that 
we need to take away. Here is another bad <?st
ring?> that needs to go too.

[download]

I need to replace all the <?string?> with <xxx> (let's say)

I have this simple little script, but of course it does not span the line like I want it to

#!/usr/bin/perl
use strict;

print "Info: ARGV is: @ARGV\n";

while (<STDIN>) {
        chomp;
        if (-e $_) { # a regular file (might be suited to your needs)
          # do something with $_ as if it were shifted from @ARGV
           print "handling file: $_\n";
           open IN, "<$_" or die "Can't open $_: $!\n";
        open OUT, ">outfile" or die "Can't open outfile: $!\n";
        while (<IN>) {
            s/<?string?>/<xxx>/g;
            print OUT;
        }
        close IN;close OUT;
        } else {
           warn "no such file: $_ \n";
        }
}

[download]

(yes, I know, I need to escape the special chars - I left them out for readability for now)

Help !

Replies are listed 'Best First'.
Re: Search and replace over line break
by almut (Canon) on Mar 12, 2010 at 16:48 UTC

    As long as you can guarantee that the substring to match will not extend over more than two lines, you could use a two-line sliding buffer that you apply the substitution to, but then split again before outputting (so you won't get everything twice): 

    #!/usr/bin/perl
use strict;
use warnings;

my $prev = "";
my $done = 0;

while (<DATA>) {

    # sliding two-line buffer
    $_ = $prev.$_;

    # get rid of middle newline, so it doesn't interfere with match
    s/\n//;

    s/<\?string\?>/<xxx>/g;

    # split buf
    # first part will be output, second part kept
    # for joining with next input line
    ($_, $prev) = unpack "a".length($prev)."a*", $_;

    print "$_\n" if 2 .. $done;

    # corner case: have last line be printed as well
    if (eof and !$done++) { $_ = ""; redo }
}

__DATA__
hello there this is a bad <?string?> that 
we need to take away. Here is another bad <?st
ring?> that needs to go too.

    [download]

    Output: 

    hello there this is a bad <xxx> that 
we need to take away. Here is another bad <xxx
> that needs to go too.

    [download]

    As you can see, line breaks may be occur anywhere in the middle of the substitution. If that's a problem, you'd have to modify the logic of splitting the sliding buffer (which can get tricky...).

    (There may be corner cases I've overlooked, but you get idea...)

[reply]
[d/l]
[select]
Re: Search and replace over line break
by Svante (Sexton) on Mar 12, 2010 at 12:34 UTC

    If there is a newline in your search string, where do you want that newline to appear in your replacement string?

    By the way, seeing as you put <> around your strings: *stern look* you are not trying to parse XML or HTML with regular expressions, are you?

[reply]
[d/l]
      I am trying to parse XML with regular expressions - not a good idea ? It's DITA actually - but that's XML also. Tell me if this is a bad idea - the issue I'm working on is fixing up some XML files that have a whole list of "bad" tags - mostly removing them, but sometime having to replace them with something else.
[reply]

        XML is not a regular language (please note that "regular" has a specific meaning with regard to a language; look it up if you do not know what this means). Therefore, you cannot generally parse it with a regular expression.

        Take an XML library that fits your need. Here are some descriptions of such.

[reply]

Back to Seekers of Perl Wisdom