Few friends have asked me how to do mass find and replace text in files under linux. There are quite a bit of options in linux to achieve mass replacing of text in files. If you are doing it file by file, you can achieve that in vi by opening, running replace and closing and going to next file. But sometimes that can be very tedious and you would rather do mass replacement on all files containing certain extension. We can do this by using sed or perl. Since most people are familiar with perl (at least most system admins and programmers), I will show you a perl way of doing it which you can use with sed as well if you wish. First step is to get perl to do what we want on one file
perl -w -i -p -e "s/search_text/replace_text/g" filename
-w turns warnings on
-i makes Perl operate on files in-place (if you would like to make backups, use -i.bak, this will save filename.bak files)
-p loops over the whole input
-e specifies Perl expression
filename works on one file at a time
Once we get the results we want, we can now pass it multiple files by doing something like:
perl -w -i -p -e "s/search_text/replace_text/g" *.php
This will search and replace within all php files in the directory you are in.
Now, let us say you want to go through your whole web directory and replace every place where you have “Perl is good” to “Perl is great”, you would use following command:
find /www_root -name "*.php"|xargs perl -w -i -p -e "s/Perl is good/perl is great/g"
find will start at /www_root, look for filenames which have .php extension, xargs takes that filename and passes it to perl as an arguement.
