[grep, sed] Find and Replace string in multiple files

Edit on GitHub
Jan 6, 2016
Last updated: May 16, 2020
Commands
4 minutes

1# replace all htttp://link to https://link in all files
2grep -rl 'http://' ./ | xargs sed -i 's|http://|https://|g'

1# if filenames have spaces, blanks or new lines
2grep -rl --null mangoes/ . | xargs -0 sed -i '' 's/mangoes/oranges/g'

why?

Here’s the scenario, i exported some files for this blog from a software that used a different image folder structure. Lots of new files with different links. So i needed a way to

find all the files spanning multiple directories containing a specific string
edit (in place) all them files found in multiple directories and replace it with new string

Basically, what i needed to do was change the path for images directory from resources/file.png to /assets/img/file.png. But i didn’t know which files had the new link structure and which files had the old one. We’re talking about 500+ files, in 25+ directories, so i can’t just go open them one by one and find stuff..

Basic Format

1grep -rl matchstring somedir/ | xargs sed -i 's/string1/string2/g'

Note: sed takes whatever follows the `s` as the separator. The forward slash `/` delimiter in the sed argument could also be a different delimiter (such as the pipe `|` character). The pipe delimiter might be useful when searching through a lot of html files if you didn't want to escape the forward slash, for instance.

matchstring is the string you want to match, e.g., “football”
string1 would ideally be the same string as matchstring, as the matchstring in the grep command will pipe only files with matchstring in them to sed.
string2 is the string that replace string1.

There may be times when you want to use grep to find only files that have some matchstring and then replace on a different string in the file than matchstring. For example, maybe you have a lot of files and only want to only replace on files that have the matchstring of ‘phonenumber’ in them, and then replace ‘555-5555’ with ‘555-1337’. Not that great of an example (you could just search files for that phone number instead of the string ‘phonenumber’), but your imagination is probably better than mine.

Example

1grep -rl 'windows' ./ | xargs sed -i 's/windows/linux/g'

This will search for the string ‘windows’ in all files relative to the current directory and replace ‘windows’ with ’linux’ for each occurrence of the string in each file.

r or --recursive is for recursive, so it searches subdirectories too
l or --files-with-matches is for listing matched files

     -l, --files-with-matches
             Only the names of files containing selected lines are written to standard output.  grep will only search a file until a match has been found, mak-
             ing searches potentially less expensive.  Pathnames are listed once per file searched.  If the standard input is searched, the string ``(standard
             input)'' is written.

     -R, -r, --recursive
             Recursively search subdirectories listed.

File names with spaces, blanks or newlines

On macOS, you may run into issues if your file names have blanks, empty spaces or newlines in them. We are using xargs to take the output of grep as our input and build the command on that. The default xargs behaviour is to delimit input with blanks and newlines. The default grep output separates filenames by newlines. See the issue?

If your file is named blah blah blah.txt, and you found it via grep, you’ll get the following error when passing it to sed via xargs

sed: ./blah : No such file or directory

xargs reads items from the standard input, delimited by blanks (which can be protected with double or single quotes or a backslash) or newlines, and executes the command (default is /bin/echo) one or more times with any initial-arguments followed by items read from standard input. Blank lines on the standard input are ignored.

Because Unix filenames can contain blanks and newlines, this default behaviour is often problematic; filenames containing blanks and/or newlines are incorrectly processed by xargs

In these situations it is better to use the -0 option, which prevents such problems. When using this option you will need to ensure that the program which produces the input for xargs also uses a null character as a separator. If that program is GNU find for example, the -print0 option does this for you.

We’re using grep to find our files and grep comes with a --null option to print a zero-byte after the file name, essentially getting rid of the newline. And to xargs we’ll pass the -0 option which changes xargs to expect NULL characters as separators, instead of spaces and newlines.

1grep -rl --null resources/ . | xargs -0 sed -i '' 's|resources/|/assets/img/|g'

Notes