Classic Shell Scripting

Command Substitution

Command substitution is the process by which the shell runs a command and replaces the command substitution with the output of the executed command. That sounds like a mouthful, but it's pretty straightforward in practice.

There are two forms for command substitution. The first form uses so-called backquotes, or grave accents (`...`), to enclose the command to be run:

for i in `cd /old/code/dir ; echo *.c`   Generate list of files in /old/code/dir
do                                       Loop over them
    diff -c /old/code/dir/$i $i | more   Compare old version to new in pager program
done

The shell first executes cd /old/code/dir ; echo *.c. The resulting output (a list of files) then becomes the list to use in the for loop.

The backquoted form is the historical method for command substitution, and is supported by POSIX because so many shell scripts exist that use it. However, all but the most simplest uses become complicated quickly. In particular, embedded command substitutions and/or the use of double quotes require careful escaping with the backslash character:

$ echo outer `echo inner1 \`echo inner2\` inner1` outer
outer inner1 inner2 inner1 outer

This example is contrived, but it illustrates how backquotes must be used. The commands are executed in this order:

1. echo inner2 is executed. Its output (the word inner2) in placed into the next command to be executed.

2. echo inner1 inner2 inner1 is executed. Its output (the words inner1 inner2 inner3) is placed into the next command to be executed.

3. Finally, echo ...