Home / server / Questions / 70939
Accepted
mezgani
Asked: 2009-10-03 14:30:04 +0800 CST

How to replace a text string in multiple files in Linux

  • 772

There are a variety of ways to replace one string of text with another across many files. Here are a few ways:

using sed and find:

sed 's/oldstring/newstring/' "$1" > "$1".new && find -iname "*.new" | sed 's/.new//' | sh

using grep and sed:

grep -rl oldstring . | xargs sed -i -e 's/oldstring/newstring/'

using grep and perl:

grep -rl oldstring . | xargs perl -pi~ -e 's/oldstring/newstring/'

Please offer your own suggestions.

linux

6 Answers

  1. Using the GNU find, xargs and sed like this: 

     find -name '*.txt' -o -name '*.html' -print0 | xargs -0 -P 1 -n 10 sed --in-place 's/oldstring/newstring/g'

    Adjust the -P and -n parameters as you like. The /g is needed so that every occurrence in a line gets replaced, not just the first one (g stands for global if I remember correctly). You can also pass a value to --in-place to make a backup.

    • 3

  2. I like perl's in-place filtering recipe.

       perl -pi.bak -e 's/from/to/' file1 file2 ...

    In context...

    % echo -e 'foo\ngoo\nboo' >test
% perl -pi.bak -e 's/goo/ber/' test
% diff -u test.bak test
--- test.bak    2010-01-06 05:43:53.072335686 -0800
+++ test    2010-01-06 05:44:03.751585440 -0800
@@ -1,3 +1,3 @@
 foo
-goo
+ber
 boo

    here is the trimmed quick-reference on the perl incantation used...

    % perl --help
Usage: perl [switches] [--] [programfile] [arguments]
  -e program        one line of program (several -e's allowed, omit programfile)
  -i[extension]     edit <> files in place (makes backup if extension supplied)
  -n                assume "while (<>) { ... }" loop around program
  -p                assume loop like -n but print line also, like sed
    • 3
  3. Best Answer

    I'd use Python for this. Put all this code into a file called mass_replace and "chmod +x mass_replace":

    #!/usr/bin/python

import os
import re
import sys

def file_replace(fname, s_before, s_after):
    out_fname = fname + ".tmp"
    out = open(out_fname, "w")
    for line in open(fname):
        out.write(re.sub(s_before, s_after, line))
    out.close()
    os.rename(out_fname, fname)


def mass_replace(dir_name, s_before, s_after):
    for dirpath, dirnames, filenames in os.walk(dir_name):
        for fname in filenames:
            f = fname.lower()
            # example: limit replace to .txt, .c, and .h files
            if f.endswith(".txt") or f.endswith(".c") or f.endswith(".h"):
                f = os.path.join(dirpath, fname)
                file_replace(f, s_before, s_after)

if len(sys.argv) != 4:
    u = "Usage: mass_replace <dir_name> <string_before> <string_after>\n"
    sys.stderr.write(u)
    sys.exit(1)

mass_replace(sys.argv[1], sys.argv[2], sys.argv[3])

    For a single search and replace of one string in one type of file, the solution with find and sed isn't bad. But if you want to do a lot of processing in one pass, you can edit this program to extend it, and it will be easy (and likely to be correct the first time).

    • 0

  4. Assuming the list of files isn't a mile long, you don't need to use xargs, as sed can handle multiple files on the command line:

    sed -i -e 's/oldstring/newstring/' `grep -rl oldstring .`
    • 0

  5. Be careful if you replace URLs with "/" character.

    An example of how to do it:

    sed -i "s%http://domain.com%http://www.domain.com/folder/%g" "test.txt"

    Extracted from: http://www.sysadmit.com/2015/07/linux-reemplazar-texto-en-archivos-con-sed.html

    • 0

  6. Thanks for some great answers everyone! This was super-helpful.

    Since I didn't have hundreds of files to replace lines in, I used a do loop, like this:

       for R in 1 2 3 4 5; do sed -i -e 's/oldstring/newstring/' file$R; done

    Hope that helps!

    • -1