Replace a string pattern in several files using the sed command and RegExp

Sometimes it is necessary to replace some text inside a file using the sed command. However, what if the text to search for is not fixed but has some pattern?

A regular expression can be used alongside sed 🙂. Let's suppose we have the following listA.txt and listB.txt files:

name=John,email=john@myurl.com
name=Eva,email=eva@myurl.com
name=Laurie,email=laurie@anotherurl.com

order=12324,email=john@myurl.com
order=45656,email=eva@myurl.com

If we want to replace the emails with a set of XXXs we could use the following shell script:

for i in list*.txt; do
    sed -i -r 's/\w+@[a-zA-Z_]+?\.[a-zA-Z]{2,3}\.?[a-zA-Z]{0,3}/XXXXX/g' $i
done

What the -r option does, is to accept as input a regular expression, instead of just searching for a fixed string. The result would be for example:

name=John,email=XXXXX
name=Eva,email=XXXXX
name=Laurie,email=XXXXX