Diff for "BashFAQ/021"

How can I replace a string with another string in a variable, a stream, a file, or in all the files in a directory?

There are a number of techniques for this. Which one to use depends on many factors, the biggest of which is what we're editing. This page also contains contradictory advice from multiple authors. This is a deeply ugly topic, and there are no universally right answers (but plenty of universally wrong ones).

Files

Before you start, be warned that editing files is a really bad idea. The preferred way to modify a file is to create a new file within the same file system, write the modified content into it, and then mv it to the original name. This is the only way to prevent data loss in the event of a crash while writing. However, using a temp file and mv means that you break hardlinks to the file (unavoidably), that you would convert a symlink to hard file, and that you may need to take extra steps to transfer the ownership and permissions (and possibly other metadata) of the original file to the new file. Some people prefer to roll the dice and accept the tiny possibility of data loss versus the greater possibility of hardlink loss and the inconvenience of chown/chmod (and potentially setfattr, setfacl, chattr...).

The other major problem you're going to face is that all of the standard Unix tools for editing files expect some kind of regular expression as the search pattern. If you're passing input you did not create as the search pattern, it may contain syntax that breaks the program's parser, which can lead to failures, or CodeInjection exploits.

Just Tell Me What To Do

If your search string or your replacement string comes from an external source (environment variable, argument, file, user input) and is therefore not under your control, then this is your best choice:

in="$search" out="$replace" perl -pi -e 's/\Q$ENV{"in"}/$ENV{"out"}/g' ./*

That will operate on all of the files in the current directory. If you want to operate on a full hierarchy (recursively), then:

in="$search" out="$replace" find . -type f -exec \
  perl -pi -e 's/\Q$ENV{"in"}/$ENV{"out"}/g' {} +

You may of course supply additional options to find to restrict which files are replaced; see UsingFind for more information.

The critical reader may note that these commands use perl which is not a standard tool. That's because none of the standard tools can do this task safely.

If you're stuck using standard tools due to a restricted execution environment, then you'll have to weigh the options below and choose the one that will do the least amount of damage to your files.

Using a file editor

The only standard tools that actually edit a file are ed and ex (vi is the visual mode for ex).

ed is the standard UNIX command-based editor. ex is another standard command-line editor. Here are some commonly-used syntaxes for replacing the string olddomain.com by the string newdomain.com in a file named file. All four commands do the same thing, with varying degrees of portability and efficiency:

## Ex
ex -sc '%s/olddomain\.com/newdomain.com/g|x' file

## Ed
# Bash
ed -s file <<< $'g/olddomain\\.com/s//newdomain.com/g\nw\nq'

# Bourne (with printf)
printf '%s\n' 'g/olddomain\.com/s//newdomain.com/g' w q | ed -s file

printf 'g/olddomain\\.com/s//newdomain.com/g\nw\nq' | ed -s file

# Bourne (without printf)
ed -s file <<!
g/olddomain\\.com/s//newdomain.com/g
w
q
!

To replace a string in all files of the current directory, just wrap one of the above in a loop:

for file in ./*; do
    [[ -f $file ]] && ed -s "$file" <<< $'g/old/s//new/g\nw\nq'
done

To do this recursively, the easy way would be to enable globstar in bash 4 (shopt -s globstar, a good idea to put this in your ~/.bashrc) and use:

# Bash 4+ (shopt -s globstar)
for file in ./**; do
    [[ -f $file ]] && ed -s "$file" <<< $'g/old/s//new/g\nw\nq'
done

If you don't have bash 4, you can use find. Unfortunately, it's a bit tedious to feed ed stdin for each file hit:

find . -type f -exec sh -c 'for f do ed -s "$f" <<!
g/old/s//new/g
w
q
!
done' sh {} +

Since ex takes its commands from the command-line, it's less painful to invoke from find:

find . -type f -exec ex -sc '%s/old/new/g|x' {} \;

Beware though, if your ex is provided by vim, it may get stuck for files that don't contain an old. In that case, you'd add the e option to ignore those files. When vim is your ex, you can also use argdo and find's {} + to minimize the amount of ex processes to run:

# Bash 4+ (shopt -s globstar)
ex -sc 'argdo %s/old/new/ge|x' ./**

# Bourne
find . -type f -exec ex -sc 'argdo %s/old/new/ge|x' {} +

You can also ask for confirmation for every replacement from A to B. You will need to type y or n every time. Please note that the A is used twice in the command. This approach is good when wrong replacements may happen (working with a natural language, for example) and the data set is small enough.

find . -type f -name '*.txt' -exec grep -q 'A' {} \; -exec vim -c '%s/A/B/gc' -c 'wq' {} \;

Using a temporary file

If shell variables are used as the search and/or replace strings, ed is not suitable. Nor is sed, or any tool that uses regular expressions. Consider using the awk code at the bottom of this FAQ with redirections, and mv.

gsub_literal "$search" "$rep" < "$file" > tmp && mv -- tmp "$file"

# Using GNU tools to preseve ownership/group/permissions
gsub_literal "$search" "$rep" < "$file" > tmp &&
  chown --reference="$file" tmp &&
  chmod --reference="$file" tmp &&
  mv -- tmp "$file"

Using nonstandard tools

sed is a Stream EDitor, not a file editor. Nevertheless, people everywhere tend to abuse it for trying to edit files. It doesn't edit files. GNU sed (and some BSD seds) have a -i option that makes a copy and replaces the original file with the copy. An expensive operation, but if you enjoy unportable code, I/O overhead and bad side effects (such as destroying symlinks), and CodeInjection exploits, this would be an option:

sed -i    's/old/new/g' ./*  # GNU, OpenBSD
sed -i '' 's/old/new/g' ./*  # FreeBSD

Those of you who have perl 5 can accomplish the same thing using this code:

perl -pi -e 's/old/new/g' ./*

Recursively using find:

find . -type f -exec perl -pi -e 's/old/new/g' {} \;   # if your find doesn't have + yet
find . -type f -exec perl -pi -e 's/old/new/g' {} +    # if it does

If you want to delete lines instead of making substitutions:

# Deletes any line containing the perl regex foo
perl -ni -e 'print unless /foo/' ./*

To replace for example all "unsigned" with "unsigned long", if it is not "unsigned int" or "unsigned long" ...:

find . -type f -exec perl -i.bak -pne \
    's/\bunsigned\b(?!\s+(int|short|long|char))/unsigned long/g' {} \;

All of the examples above use regular expressions, which means they have the same issue as the sed code earlier; trying to embed shell variables in them is a terrible idea, and treating an arbitrary value as a literal string is painful at best.

If the inputs are not under your direct control, you can pass them as variables into both search and replace strings with no unquoting or potential for conflict with sigil characters:

in="$search" out="$replace" perl -pi -e 's/\Q$ENV{"in"}/$ENV{"out"}/g' ./*

Or, wrapped in a useful shell function:

# Bash
# usage: replace FROM TO [file ...]
replace() {
  local in="$1" out="$2"; shift 2
  in="$in" out="$out" perl -p ${1+-i} -e 's/\Q$ENV{"in"}/$ENV{"out"}/g' "$@"
}

This wrapper passes perl's -i option if there are any filenames, so that they are "edited in-place" (or at least as far as perl does such a thing -- see the perl documentation for details).

Variables

If you want to replace content within a variable, this can (and should) be done very simply with Bash's parameter expansion:

var='some string'; search=some; rep=another

# Bash
var=${var//"$search"/$rep}

It's a lot harder in POSIX:

# POSIX function

# usage: string_rep SEARCH REPL STRING
# replaces all instances of SEARCH with REPL in STRING
string_rep() {
  # initialize vars
  in=$3
  unset out

  # SEARCH must not be empty
  case $1 in '') return; esac

  while
    # break loop if SEARCH is no longer in "$in"
    case "$in" in
      *"$1"*) ;;
      *) break;;
    esac
  do
    # append everything in "$in", up to the first instance of SEARCH, and REP, to "$out"
    out=$out${in%%"$1"*}$2
    # remove everything up to and including the first instance of SEARCH from "$in"
    in=${in#*"$1"}
  done

  # append whatever is left in "$in" after the last instance of SEARCH to out, and print
  printf '%s%s\n' "$out" "$in"
}

var=$(string_rep "$search" "$rep" "$var")

# Note: POSIX does not have a way to localize variables. Most shells (even dash and 
# busybox), however, do. Feel free to localize the variables if your shell supports
# it. Even if it does not, if you call the function with var=$(string_rep ...), the
# function will be run in a subshell and any assignments it makes will not persist.

In the bash example, the quotes around "$search" prevent the contents of the variable to be treated as a shell pattern (also called a glob). Of course, if pattern matching is intended, do not include the quotes. If "$rep" were quoted, however, the quotes would be treated as literal.

Parameter expansions like this are discussed in more detail in Faq #100.

Streams

If you wish to modify a stream, and if your search and replace strings are known in advance, then use the stream editor:

some_command | sed 's/foo/bar/g'

sed uses regular expressions. In our example, foo and bar are literal strings. If they were variables (e.g. user input), they would have to be rigorously escaped in order to prevent errors. This is very impractical, and attempting to do so will make your code extremely prone to bugs. Embedding shell variables in sed commands is never a good idea -- it is a prime source of CodeInjection bugs.

You could also do it in Bash itself, by combining a parameter expansion with Faq #1:

search=foo rep=bar

while IFS= read -r line; do
  printf '%s\n' "${line//"$search"/$rep}"
done < <(some_command)

# or

some_command | while IFS= read -r line; do
  printf '%s\n' "${line//"$search"/$rep}"
done

If you want to do more processing than just a simple search/replace, this may be the best option. Note that the last example runs the loop in a SubShell. See Faq #24 for more information on that.

You may notice, however, that the bash loop above is very slow for large data sets. So how do we find something faster, that can replace literal strings? Well, you could use awk. The following function replaces all instances of STR with REP, reading from stdin and writing to stdout.

# usage: gsub_literal STR REP
# replaces all instances of STR with REP. reads from stdin and writes to stdout.
gsub_literal() {
  # STR cannot be empty
  [[ $1 ]] || return

  # string manip needed to escape '\'s, so awk doesn't expand '\n' and such
  awk -v str="${1//\\/\\\\}" -v rep="${2//\\/\\\\}" '
    # get the length of the search string
    BEGIN {
      len = length(str);
    }

    {
      # empty the output string
      out = "";

      # continue looping while the search string is in the line
      while (i = index($0, str)) {
        # append everything up to the search string, and the replacement string
        out = out substr($0, 1, i-1) rep;

        # remove everything up to and including the first instance of the
        # search string from the line
        $0 = substr($0, i + len);
      }

      # append whatever is left
      out = out $0;

      print out;
    }
  '
}

some_command | gsub_literal "$search" "$rep"


# condensed as a one-liner:
some_command | awk -v s="${search//\\/\\\\}" -v r="${rep//\\/\\\\}" 'BEGIN {l=length(s)} {o="";while (i=index($0, s)) {o=o substr($0,1,i-1) r; $0=substr($0,i+l)} print o $0}'

CategoryShell