What if you wanted to find and replace an instance or all instances of a string within a file from the command line? “Why would I want to do that?”, you may ask – because we’re too lazy to open a text/code editor and because it’s kewl! You would use a command called sed (StreamEDitor) thusly:

1

sed -i 's/the_string_to_find/the_string_to_replace/g' the_file.txt
  • i – edit in place/save changes back to the file
  • s – substitute
  • /find/replace/g – regular expression matching with the global flag

Example:

1

sed -i 's/server\.js/express\.js/g' package.json

Using extended regex matching

You can use the -r flag when you want to use extended regexs. So in order to match and remove whitespace you could use the following:

1

sed -r 's/\s+//g' <filename>

And you’re off to the races :)