Using Quotes in Shell Scripts

* Double quote (?): Variables are expanded when enclosed in double quotes

* Single quote (?): Variables within single quotes are not expanded

* Backward quote (`): Commands within backward quotes are executed and their output substituted into that location.

The following examples should help make the differences between the types of quotes more evident.

$ myname='Terry Clark'
$ echo 'My name is $myname'
My name is $myname
$ echo "My name is $myname"
My name is Terry Clark
$ echo 'The name of this computer is `hostname`'
The name of this computer is `hostname`
$ echo "The name of this computer is `hostname`"
The name of this computer is appsvr.mytec.com

So we see that when enclosed in single quotes text is treated very literally, but when enclosed in double quotes variables and commands in backward quotes are evaluated and their values output.

Exit Status

Whenever a command or shell script completes successfully, it sets a hidden status code of zero. If the command is unsuccessful, it sets a nonzero hidden status code.  This completion status code is known as the exit status.  The exit status of the last command or script that was run is contained in the special shell variable, $?.

Most of the time we never look at this value and instead check to see if the command did what we want or look for errors in the output of commands.  In a shell script, however, we may want to check the exit status to make sure everything is going OK.  The exit status of the last command can be displayed as follows:

$ ls
example1.fil  example2.xxx  examples  test.bsh  umask_example.fil
$ echo $?
0
$ ls *.txt
ls: *.txt: No such file or directory
$ echo $?
1

The conditional statements shown later in this section can be used to make decisions based on the exit status of the previous command.  This can be useful for handling and reporting errors in a shell script.

The read Statement

The read command is used to accept input from standard input such as a keyboard and store it into variables.  This is useful if you want to prompt for input from the user running your script.

The read command reads one line from the standard input and the first word is assigned to the first variable, the second word to the second variable, and so on, with leftover words assigned to the last variable.  If there are fewer words read from the input stream than variables, the remaining variables are assigned empty values.

Here are some examples of the read command:

$ read name1 name2
Terry Clark
$ echo $name2
Clark
$ echo "Name1: $name1    Name2: $name2"
Name1: Terry    Name2: Clark
$ read fullname
Terry L. Clark
$ echo $fullname
Terry L. Clark

The ?p option can be used to specify a prompt to be displayed to the user.  The prompt can be used to request that the user enter specific information.  The following is an example of using read with the ?p option to prompt the user for input:

$ read -p "Enter your name please: "  fullname

Now you can start to see how we can make an interactive shell script which prompts the user for input.  Sometimes we want to collect information when the shell script is started rather than once it is running.  For that type of information command line arguments can be used.

Command Line Arguments

Arguments or parameters can be passed to the script from the command line that invoked the script just like we have been passing arguments to commands.  Arguments passed to the script are available in special shell variables which are referenced by number.  The first argument can be referenced using the variable $1, the second argument using $2, etc.  The $0 variable is special and contains the name of the command used to invoke the script or command name.  The following is an example of using command line arguments.

#
#  This script accepts a user?s first and last name and
#  displays the current login username, hostname, and home
#  directory.
#

Note that anything after a pound sign (#) in a shell script is a comment and will be ignored when the script is run.

If we were to run the above script here is what you would get as a result:

$ ./mylogin Terry Clark

We can see how the script has captured the input and substituted the $0, $1 and $2 variables into the context of the output.  We have also used some of the global shell variables discussed earlier like $USER, $HOSTNAME and $HOME.

Sometimes the input you're looking for won't come from a user at a keyboard but rather from a file or from the output of another command.  Other times you may want to output information to an output file or a log.  Next we'll look at how to do these things within your shell commands and scripts.

Redirection of Standard Input & Output

In Linux, the standard input device is the keyboard, and the standard output device is the display screen.  These are the defaults for input and output; however, Linux does allow the redirection of both standard input and output to files.  In other words, Linux accepts input from files and can direct output to files.

Three symbols are used to indicate redirection to Linux.  Table 8.2 below shows the meanings of the symbols:

Table 8.2: Redirection symbols and their meanings

The following is an example of redirecting command output to a file using the > symbol.  We then examine the file with the cat command.

$ ls -alt > dir.lst
$ cat dir.lst
total 68
drwx------    4 tclark   tclark       4096 Feb 24 23:27 .
-rw-rw-r--    1 tclark   tclark          0 Feb 24 23:27 dir.lst
-rw-------    1 tclark   tclark      11629 Feb 24 21:24 .bash_history
-rwxrwxr-x    1 tclark   tclark        325 Feb 24 21:17 mylogin
-rw-------    1 tclark   tclark       4151 Feb 24 21:17 .viminfo
-rwxrwxr-x    1 tclark   tclark        208 Feb 19 22:52 test.bsh
drwxrwxr-x    2 tclark   authors      4096 Feb  3 23:11 examples
-rw-rw-r--    1 tclark   tclark          0 Feb  3 09:00 example1.fil
-rw-r--r--    1 tclark   authors         0 Jan 27 00:22 umask_example.fil
drwxr-xr-x    8 root     root         4096 Jan 25 22:16 ..
-rw-rw-r--    1 tclark   tclark          0 Jan 13 21:13 example2.xxx
-rw-r--r--    1 tclark   tclark        120 Aug 24  2004 .gtkrc
-rw-r--r--    1 tclark   tclark         24 Aug 18  2004 .bash_logout
-rw-r--r--    1 tclark   tclark        191 Aug 18  2004 .bash_profile
-rw-r--r--    1 tclark   tclark        124 Aug 18  2004 .bashrc
-rw-r--r--    1 tclark   tclark        237 May 22  2003 .emacs
-rw-r--r--    1 tclark   tclark        220 Nov 27  2002 .zshrc
drwxr-xr-x    3 tclark   tclark       4096 Aug 12  2002 .kde

If instead we want to add on to the end of a file the >> symbol can be used.  Here we are placing the current time at the end of the time.txt file then examining the file with cat:

$ date >> time.txt
$ date >> time.txt
$ date >> time.txt
$ cat time.txt

Here is how a file can be redirected into a command.  In this case we're using the dir.lst file from the example above.

$ less < dir.lst
total 68
drwx------    4 tclark   tclark       4096 Feb 24 23:27 .
-rw-rw-r--    1 tclark   tclark          0 Feb 24 23:27 dir.lst
-rw-------    1 tclark   tclark      11629 Feb 24 21:24 .bash_history
-rwxrwxr-x    1 tclark   tclark        325 Feb 24 21:17 mylogin
-rw-------    1 tclark   tclark       4151 Feb 24 21:17 .viminfo
-rwxrwxr-x    1 tclark   tclark        208 Feb 19 22:52 test.bsh
drwxrwxr-x    2 tclark   authors      4096 Feb  3 23:11 examples
-rw-rw-r--    1 tclark   tclark          0 Feb  3 09:00 example1.fil
-rw-r--r--    1 tclark   authors         0 Jan 27 00:22 umask_example.fil
drwxr-xr-x    8 root     root         4096 Jan 25 22:16 ..
-rw-rw-r--    1 tclark   tclark          0 Jan 13 21:13 example2.xxx
-rw-r--r--    1 tclark   tclark        120 Aug 24  2004 .gtkrc
-rw-r--r--    1 tclark   tclark         24 Aug 18  2004 .bash_logout
-rw-r--r--    1 tclark   tclark        191 Aug 18  2004 .bash_profile
-rw-r--r--    1 tclark   tclark        124 Aug 18  2004 .bashrc
-rw-r--r--    1 tclark   tclark        237 May 22  2003 .emacs
-rw-r--r--    1 tclark   tclark        220 Nov 27  2002 .zshrc
drwxr-xr-x    3 tclark   tclark       4096 Aug 12  2002 .kde
(END)

Sometimes we want to cut out the middle man, avoid using files, and redirect output directly from one command to another.  That's where the pipe comes in.

Pipes

The pipe can be used to pass the output from one command or program to another command or program without creating an intermediate file.  The use of pipes is a convenient way to string together multiple commands on a command line.  The pipe symbol is the vertical bar (|), which is placed between commands to invoke a pipe.

To see the files in the current working directory that have .fil as the file name extension, a pipe could be formed between the ls and grep commands as follows:

$ ls -l|grep .fil
-rw-rw-r--    1 tclark   tclark          0 Feb  3 09:00 example1.fil
-rw-r--r--    1 tclark   authors         0 Jan 27 00:22 umask_example.fil

It's important to note that the grep command here is not only looking at the file name but is actually looking for the string .fil anywhere in the lines of the ls ?l output.

The following more complicated pipe example show the current processes (ps ?ef), redirects the list of processes to filter for only ones with root in the output (grep root), and counts the number of remaining lines with the word count command (wc ?l).  The resulting string of commands returns the total number of current root processes.

$ ps -ef|grep root|wc -l