Introduction to Shell Scripting

Lecture
1. Shell scripts are small programs. They let you automate multi-step processes, and
give you the capability to use decision-making logic and repetitive loops.
2. Although we have been using the tcsh shell for our interactive work, we will
use the Bourne shell (sh) for scripts. Most UNIX scripts are written in some
variant of the Bourne shell.
3. Consider this sample script (you’ve seen it before in the PATH example). It can
be found at ~unixinst/bin/progress.
#! /bin/sh
#
# Sample shell script for Introduction to UNIX class.
# Jason R. Banfelder.
# Displays basic system information and UNIX
# students' disk usage.
#
# Show basic system information
echo `hostname`: `w | head -1`
echo `who | cut -d" " -f1 | sort | uniq | \
grep "^unixst" | wc -l` students are logged in.
#
# Generate a disk usage report.
echo "-----------------"
echo "Disk Usage Report"
echo "-----------------"
cd /home
# Loop over each student's home directory...
for STUDENT_ID in `ls -1d unixst*`
do
# ...and show how much disk space is used.
du -skL /home/$STUDENT_ID
done
a. All scripts should begin with a ‘shebang’ (#!) to give the name of the
shell.
b. Comments begin with a hash.
c. You can use all of the UNIX commands you know in your scripts.
d. The results of commands in backwards quotes (upper-left of your
keyboard) are substituted into commands.
e. Use a $ before variable names to use the value of that variable.
f. Note the for loop construction. See the Introduction to UNIX manual
for details on the syntax.
4. Scripts have to be executed, so you need to chmod the script file. Use ls –l
to see the file modes.

1
Exercise
1. Write a script to print out the quotations from each directory.
a. Did you write the script from scratch, or copy and modify the example
above?
2. Create a subdirectory called bin in your home directory. Move your script
there.
3. Permanently add your bin directory to your PATH.
a. It is a UNIX tradition to put your useful scripts and programs into a
directory named bin.
4. Save your script’s output to a file, and e-mail the file to yourself.
a. mail [email protected] < AllMyQuotations.txt
b. We hope you enjoy this list of quotations as a souvenir of this class.

2
More Scripting Techniques
Lecture
1. As you write scripts, you will find you want to check for certain conditions before
you do things. For example, in the script from the previous exercise, you don’t
want to print out the contents of a file unless you have permission to read it.
Checking this will prevent warning messages from being generated by your
scripts.
a. The following script fragment checks the readability of a file. Note that
this is a script fragment, not a complete script. It won’t work by itself
(why not?), but you should be able to incorporate the idea into your own
scripts.
if [ -r $STUDENT_ID/quotation ]; then
echo
cat $STUDENT_ID/quotation
fi
b. Note the use of the if…fi construct. See the Introduction to UNIX
manual that accompanies this class for more details, including using
else blocks.
i. In particular, note that you must have spaces inside the brackets in
the test expression.
c. Note how the then command is combined on the same line as the if
statement by using the ; operator.
d. You can learn about many other testing options (like –r) by reading the
results of the man test command.
2. The read command is useful for reading input (either from a file or from an
interactive user at the terminal) and assigning the results to variables.
#! /bin/sh
#
# A simple start at psychiatry.
# (author to remain nameless)
echo "Hello there."
echo "What is your name?"
read PATIENT_NAME
echo "Please have a seat, ${PATIENT_NAME}."
echo "What is troubling you?"
read PATIENT_PROBLEM
echo -n "Hmmmmmm.... '"
echo -n $PATIENT_PROBLEM
echo "' That is interesting... Tell me more..."
a. Note how the variable name is in braces. Use braces when the end of the
variable name may be ambiguous.

3
 3. The read command can also be used in a loop to read one line at a time from a
file.
while read line; do
echo $line
<your script code here>
done < input.txt
a. You can also use [] tests as the condition for the loop to continue or
terminate in while commands.
b. Also see the until command for a similar loop construct,
4. You can also use arguments from the command line as variables.
while read line; do
echo $line
<more script code here>
done < $1
a. $1 is the first argument after the command, $2 is the second, etc.

Exercise
1. Modify your quotation printing script to test the readability of files before trying
to print them.

4
Scripting Expressions
Lecture
1. You can do basic integer arithmetic in your scripts.
a. total=`expr $1 + $2 + $3 + $4 + $5` will add the first five
numerical arguments to the script you are running, and assign them to the
variable named total.
b. Note the use of the backwards quotes.
i. Try typing expr 5 + 9 at the command line. Recall, the
backwards quotes substitute the result of a command into your
script.
ii. What happens if one of the arguments is not a number?
c. Some operators, such as multiplication and parentheses for grouping, need
to be escaped out because of their special meaning to the shell.
i. gccontentMin=
`expr \( $gcMin \* 100 \) / $total`
2. You can also do greater than and less than comparisons, logic, etc. with the
expr command.
a. You guessed it. Read the man page.
b. There are also useful operators for working with string variables. In
particular, see the “Expression Evaluation” section of the Introduction to
UNIX manual to learn about the match(:), substr, index, and
length operators.

Exercise
1. When we learned about csplit, we saw that we had to know how many
sequences were in a .fasta file to properly construct the command.
a. Write a script to do this work for you.
#! /bin/sh
#
# Intelligently split a fasta file containing
# multiple sequences into multiple files each
# containing one sequence.
#
seqcount=`grep -c '^>' $1`
echo "$seqcount sequences found."
if [ $seqcount -le 1 ]; then
echo "No split needed."
exit
elif [ $seqcount -eq 2 ]; then
csplit -skf seq $1 '%^>%' '/^>/'
else
repcount=`expr $seqcount - 2`
csplit -skf seq $1 '%^>%' '/^>/' \{${repcount}\}
fi

5
 b. Expand this script to rename each of the resultant files to reflect the
sequence’s GenBank ID.
>gi|37811772|gb|AAQ93082.1| taste receptor T2R5 [Mus musculus]
This is shown underlined and in italics in the example above.
i. How would you handle fasta headers without a GenBank ID?
c. Expand your script to sort the sequence files into two directories, one for
nucleotide sequences (which contain primarily A, T, C, G), and one for
amino acid sequences.
i. How would you handle situations where the directories do/don’t
already exist?
ii. How would you handle situations where the directory name
already exists as a file?
iii. When does all this checking end???
2. What does this script do?
#! /bin/sh
gcounter=0
ccounter=0
tcounter=0
acounter=0
ocounter=0
while read line ; do
isFirstLine=`echo "$line" | grep -c '^>'`
if [ $isFirstLine -ne 1 ]; then
lineLength=`echo "$line" | wc -c`
until [ $lineLength -eq 1 ]; do
base=`expr substr "$line" 1 1`
case $base in
"a"|"A")
acounter=`expr $acounter + 1`
;;
"c"|"C")
ccounter=`expr $ccounter + 1`
;;
"g"|"G")
gcounter=`expr $gcounter + 1`
;;
"t"|"T")
tcounter=`expr $tcounter + 1`
;;
*)
ocounter=`expr $ocounter + 1`
;;
esac
line=`echo "$line" | sed 's/^.//'`
lineLength=`echo "$line" | wc -c `
done
fi
done < $1
echo $gcounter $ccounter $tcounter $acounter $ocounter
3. Write a script to report the fraction of GC content in a given sequence.
a. How can you use the output of the above script to help you in this?
6