Introducing UNIX and Linux

The command to accomplish this is test. Followed by arguments, test will give an exit status of 0 if the arguments evaluate to True. There are two ways of invoking test:

and we shall use the latter for the rest of this book. To give you the flavour, the following checks whether file testfile exists and displays a suitable message if it does:

The operator -e, when presented as an argument to test, examines the following argument and, if that file exists, the command succeeds (exit status 0), otherwise it fails with exit status 1. Various options are available to test.

Worked example

Write a script which will read in the name of a file and print out a message indicating whether or not it is a directory.
Solution: Use test with option -d to check the file, and || and && to control which message is output.

echo Input a file name:   # Prompt the user ...
read FILENAME             # input a file name ...
([ -d $FILENAME ] &&      # check it's a directory ...
                          # then confirm this if so
       echo $FILENAME is a directory) ||
   echo $FILENAME is not a directory

A regular file is essentially one that is not a directory; there are other sorts of non-regular files, such as FIFO files, but they do not concern us here.

For instance, to check whether variable NAME has been set a value that is not the null string, we might have:

Note that we have enclosed $NAME in double quotes; test expects to get an argument, and if we did not enclose it in quotes, and NAME was unset (or contained only whitespace), the line would become

Worked example

Write a script to greet the person running it if they are logged on as user chris.
Solution: Use logname to check the user's name (not the variable LOGNAME, which might accidentally have been changed), and test to compare it with chris.

[ "$( logname )" = chris ] && echo Hello Chris

The numerical checks listed that test can perform are the principal way of doing numerical comparisons using the shell. They only work with whole numbers, however, and if you wish to perform complex tasks using floating-point numbers you are advised to use bc. As an example, we code the solution to the question posed at the start of this section, which was: 'If file A is smaller than 100 lines then display it on the terminal, otherwise tell me that it's bigger than 100 lines.'

Worked example

Write a script to request you to type in a number, and then to guess its square; it should then either congratulate you or tell you the correct answer.
Solution: After reading in a number into variable NUMBER, construct an expression $NUMBER * $NUMBER to be piped to bc, assigning the output of the calculation to SQUARE. Then, after reading in the user's guess into variable GUESS, use test to check whether GUESS and SQUARE are the same.

Notice the different 'style' of comments used here - each on a line of its own; this style is preferred with long command lines.

# Prompt the user and read in the number
echo Type in a number:
read NUMBER

# Evaluate the square of the number using bc
SQUARE=$( echo "$NUMBER * $NUMBER" | bc )

# Prompt the user and read in the guessed answer
echo Guess its square:
read GUESS

# If the guess is equal to the square, confirm ...
([ "$GUESS" -eq "$SQUARE" ] && echo Correct) ||
# otherwise display the correct answer
    echo The correct answer is $SQUARE

Since test requires arguments, you must separate those arguments from the word test by whitespace. Similarly, if you are using the square bracket notation for test, you should separate the square brackets from what is inside the brackets. Otherwise, the brackets themselves would become part of the strings which they should enclose. For instance,

would attempt to compare string hello with string hello] and think that you had forgotten to provide the closing square bracket.

Introducing UNIX and Linux

Introduction to shells

The 'test' statement

Worked example

Worked example

Worked example