Replacing Strings in Files - Page 3

Now that we have a basic handle on creating regular expressions to match different variations of strings, it's time to introduce the search and replace operator and write a script to perform a transformation on a file. The search and replace operator is placed in front of a set of forward slashes. The difference is that there is a search expression and a replace expression:

s/search/replace/

Continuing from our set of examples in the previous section, let's pretend that we need to replace my name as maintainer of the Web site. Instead, Bob will be filling in for me so we need to replace all occurrences of Jonathan with Bob. We'll use the regular expression we built in the last section as the search string:

s/Jon(ath[ao]n)?/Bob/gi

Notice that we used the g and i modifiers again. They work the same way in a search and replace string in that they modify the search part of the operation. Next we need to write a script that we can run from the command line that will run the search and replace expression against a file.

1 my $file = $ARGV[0];
2 open (FILE,$file) || die "Cannot read from $file";
3 open (TMP, ">$file.$$") || die "Cannot write to $file.$$\n";
4 my $counter = 0;
5
6 while (<FILE>) {
7 $counter++ if s/Jon(ath[ao]n)?/Bob/gi;
8 print TMP;
9 }
10
11 close FILE,TMP;
12 print "Found Jonathan $counter times\n";
13 rename "$file.$$",$file || die "Cannot update $file\n";

Let's go through the script line-by-line so that we understand what's happening. In line 1, we're grabbing the filename that was passed on the command line. If we named the script match.pl and we wanted to run it against index.html, the command line syntax would be:

match.pl index.html

On a Unix box, you could either place the location of the Perl binary at the top of the script or prepend the argument with Perl, perl match.pl index.html. On Windows, if you have ActivePerl from ActiveState Corp. installed on your box, it will know to run the script with the Perl interpreter automatically based upon the .pl file extension.

In lines 2 and 3, we're opening the file that was passed on the command line and also opening a temporary file, which will be used to write a copy of the file with the changes. On line 4, we initialize the $counter variable that will be used to count the occurrences of the match for reference. In line 6, we're using a while statement to loop over each line in the file, one at a time. Placing the FILE filehandle inside of the less-than and greater-than characters automatically reads one line into the $_ variable, which is a special variable in Perl. This is equivalent to saying:

while ($_ = <FILE>) {

In line 7, we search the $_ string with the search and replace operator. If a match is found, the match is replaced with Bob and the $counter variable is incremented by one. In line 8, we print the line to our temporary file. Again, we're using the magical $_ variable. By not specifying a variable, Perl automatically prints $_, which contains our string read from our file in the while loop. This is equivalent to saying:

print TMP $_;

In line 11, we close the filehandles for the input and temporary files. Line 12 reports the number of occurrences of my name that were found in the file. Finally, line 13 moves the temporary file to the file that was specified on the command line, replacing the old file with Jonathan in it with the new file that contains Bob instead of Jonathan.

Now we have a simple script that can be modified to change strings in files. By now you should have a handle on the usefulness and syntax of regular expressions in Perl. You might even be able to save some time using the script to do search and replaces on files instead of doing it by hand. But we still haven't harnessed the power of Perl. We'll do that next.

Regular Expressions Introduced - Page 2
Weaving Magic With Regular Expressions
Replacing Strings in Multiple Files - Page 4