A Tutorial for Solving Real World Problems with Bash Scripts

Traducciones al Español
Estamos traduciendo nuestros guías y tutoriales al Español. Es posible que usted esté viendo una traducción generada automáticamente. Estamos trabajando con traductores profesionales para verificar las traducciones de nuestro sitio web. Este proyecto es un trabajo en curso.
Create a Linode account to try this guide with a $ credit.
This credit will be applied to any valid services used during your first  days.

Introduction

This guide presents some of the advanced capabilities of the bash shell by showing practical and fully functional bash scripts. It also illustrates how you can work with dates and times in bash scripts and how to write and use functions in bash.

In This Guide

In this guide, you will find the following information about bash scripts:

Functions in bash shell

The bash scripting language has support for functions. The parameters of a function can be accessed as $1, $2, etc. and you can have as many parameters as you want. If you are interested in finding out the name of the function, you can use the FUNCNAME variable. Functions are illustrated in functions.sh, which is as follows:

File: functions.sh
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
#!/bin/bash

function f1 {
    echo Hello from $FUNCNAME!
    VAR="123"
}

f2() {
    p1=$1
    p2=$2
    sum=$((${p1} + ${p2}))
    echo "${sum}"
}

f1
echo ${VAR}

mySum="$(f2 1 2)"
echo mySum = $mySum

mySum="$(f2 10 -2)"
echo mySum = $mySum

Run the script with the following command:

./functions.sh

The output will look like this:

Hello from f1!
123
mySum = 3
mySum = 8
Note

If you want to check whether a function parameter exists or not, you can use the statement:

if [ -z "$1" ]

Using bash Functions as Shell Commands

This is a trick that allows you to use bash functions as shell commands. You can execute the above code as

. ./functions.sh

Notice the dot in front of the text file. After that you can use f1 as a regular command in the terminal where you executed . ./my_function.sh. You will also be able to use the f2 command with two integers of your choice to quickly calculate a sum. If you want that function to be globally available, you can put its implementation to a bash configuration file that is automatically executed by bash each time a new bash session begins. A good place to put that function implementation would be ~/.bash_profile.

Working with Dates and Times

Bash allows you to work with dates and times using traditional UNIX utilities such as date(1). The main difficulty many programmers run into when working with dates and times is getting or using the correct format. This is a matter of using date(1) with the correct parameters and has nothing to do with bash scripting per se. Using date(1) as date +[something] means that we want to use a custom format – this is signified by the use of + in the command line argument of date(1).

A good way to create unique filenames is to use UNIX epoch time or, if you want your filename to be more descriptive, a date-time combination. The unique nature of the filename is derived from a focus on a higher level of detail in defining your output. If done correctly, you will never have the exact same time value even if you execute the script multiple times on the same UNIX machine.

The example that follows will shed some light on the use of date(1).

Using Dates and Times in bash scripts

The code of dateTime.sh is the following:

File: dateTime.sh
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
#!/bin/bash

# Print default output
echo `date`

# Print current date without the time
echo `date +"%m-%d-%y"`

# Use 4 digits for year
echo `date +"%m-%d-%Y"`

# Display time only
echo `date +"%T"`

# Display 12 hour time
echo `date +"%r"`

# Time without seconds
echo `date +"%H:%M"`

# Print full date
echo `date +"%A %d %b %Y %H:%M:%S"`

# Nanoseconds
echo Nanoseconds: `date +"%s-%N"`

# Different timezone by name
echo Timezone: `TZ=":US/Eastern" date +"%T"`
echo Timezone: `TZ=":Europe/UK" date +"%T"`

# Print epoch time - convenient for filenames
echo `date +"%s"`

# Print week number
echo Week number: `date +"%V"`

# Create unique filename
f=`date +"%s"`
touch $f
ls -l $f
rm $f

# Add epoch time to existing file
f="/tmp/test"
touch $f
mv $f $f.`date +"%s"`
ls -l "$f".*
rm "$f".*

If you want an even more unique filename, you can also use nanoseconds when defining the behaviour of your script.

Run the dateTime script:

./dateTime.sh

The output of dateTime.sh will resemble the following:

Fri Aug 30 13:05:09 EST 2019
08-30-19
08-30-2019
13:05:09
01:05:09 PM
13:05
Friday 30 Aug 2019 13:05:09
Nanoseconds: 1567159562-373152585
Timezone: 06:05:09
Timezone: 10:05:09
1567159509
Week number: 35
-rw-r--r--  1 mtsouk  staff  0 Aug 30 13:05 1567159509
-rw-r--r--  1 mtsouk  wheel  0 Aug 30 13:05 /tmp/test.1567159509

Bash scripts for Administrators

This section will present some bash scripts that are generally helpful for UNIX system administrators and power users.

Watching Free Disk Space

The bash script that follows watches the free space of your hard disks and warns you when that free space drops below a given threshold – the value of the threshold is given by the user as a command line argument. Notice that if the program gets no command line argument, a default value is used as the threshold.

File: freeDisk.sh
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
#!/bin/bash

# default value to use if none specified
PERCENT=30

# test for command line arguement is present
if [[ $# -le 0 ]]
then
    printf "Using default value for threshold!\n"
# test if argument is an integer
# if it is, use that as percent, if not use default
else
    if [[ $1 =~ ^-?[0-9]+([0-9]+)?$ ]]
    then
        PERCENT=$1
    fi
fi

let "PERCENT += 0"
printf "Threshold = %d\n" $PERCENT

df -Ph | grep -vE '^Filesystem|tmpfs|cdrom' | awk '{ print $5,$1 }' | while read data;
do
    used=$(echo $data | awk '{print $1}' | sed s/%//g)
    p=$(echo $data | awk '{print $2}')
    if [ $used -ge $PERCENT ]
    then
        echo "WARNING: The partition \"$p\" has used $used% of total available space - Date: $(date)"
    fi
done
  • The sed s/%//g command is used for omitting the percent sign from the output of df -Ph.
  • df is the command to report file system disk space usage, while the options -Ph specify POSIX output and human-readable, meaning, print sizes in powers of 1024.
  • awk(1) is used for extracting the desired fields from output of the df(1) command.

Run ./freeDisk.sh with this command:

./freeDisk.sh

The output of freeDisk.sh will resemble the following:

Using default value for threshold!
Threshold = 30
WARNING: The partition "/dev/root" has used 61% of total available space - Date: Wed Aug 28 21:14:51 EEST 2019
Note
This script and others like it can be easily executed as cron jobs and automate tasks the UNIX way.

Notice that the code of freeDisk.sh looks relatively complex. This is because bash is not good at the conversion between strings and numeric values – more than half of the code is for initializing the PERCENT variable correctly.

Rotating Log Files

The presented bash script will help you to rotate a log file after exceeding a defined file size. If the log file is connected to a server process, you might need to stop the process before the rotation and start it again after the log rotation is complete – this is not the case with rotate.sh.

File: rotate.sh
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
#!/bin/bash

f="/home/mtsouk/connections.data"

if [ ! -f $f ]
then
  echo $f does not exist!
  exit
fi

touch ${f}
MAXSIZE=$((4096*1024))

size=`du -b ${f} | tr -s '\t' ' ' | cut -d' ' -f1`
if [ ${size} -gt ${MAXSIZE} ]
then
    echo Rotating!
    timestamp=`date +%s`
    mv ${f} ${f}.$timestamp
    touch ${f}
fi
  • Note that the path to the log file /home/mtsouk/connections.data will not exist by default. You’ll need to either use a log file that already exists like kern.log on some Linux systems, or replace it with a new one.

  • Additionally, the value of MAXSIZE can be a value of your choice, and the script can be edited to suit the needs of your own configuration – you can even make changes to the existing code and provide the MAXSIZE value as a command line argument to the program.

  • The du command is used to estimate the file space usage. It’s use to track the files and directories that are consuming excessive space on the hard disk. The -b option tells this command to print the size in bytes.

Run the rotate script with the following command:

./rotate.sh

The output of rotate.sh when it has reached the threshold defined by MAXSIZE will resemble the following:

Rotating!

After running, two files will be created on the system. You can see them with this command:

ls -l connections.data*
-rw-r--r-- 1 mtsouk mtsouk       0 Aug 28 20:18 connections.data
-rw-r--r-- 1 mtsouk mtsouk 2118655 Aug 28 20:18 connections.data.1567012710

If you want to make rotate.sh more generic, you can provide the name of the log file as a command line argument to the bash script.

Monitoring the Number of TCP Connections

The presented bash script calculates the number of TCP connections on the current machine and prints that on the screen along with date and time related information.

File: tcpConnect.sh
1
2
3
4
5
6
#!/bin/bash

C=$(/bin/netstat -nt | tail -n +3 | grep ESTABLISHED | wc -l)
D=$(date +"%m %d")
T=$(date +"%H %M")
printf "%s %s %s\n" "$C" "$D" "$T"
  • The main reason for using the full path of netstat(1) when calling it is to make the script as secure as possible.
  • If you do not provide the full path then the script will search all the directories of the PATH variable to find that executable file.
  • Apart from the number of established connections (defined by the C variable), the script prints the month, day of the month, hour of the day, and minutes of the hour. If you want, you can also print the year and seconds.

Execute the tcpConnect script with the following command:

./tcpConnect.sh

The output will be similar to the following:

8 08 28 16 22

tcpConnect.sh can be easily executed as a cron(8) by adding the following to your cron file:

*/4 * * * * /home/mtsouk/bin/tcpConnect.sh >> ~/connections.data

The previous cron(8) job executes tcpConnect.sh every 4 minutes, every hour of each day and appends the results to ~/connections.data in order to be able to watch or visualize them at any time.

Additional Examples

Sorting in bash

The presented example will show how you can sort integer values in bash using the sort(1) utility:

File: sort.sh
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
#!/bin/bash

# test that at least one argument was passed
if [[ $# -le 0 ]]
then
    printf "Not enough arguments!\n"
    exit
fi

count=1

for arg in "$@"
do
    if [[ $arg =~ ^-?[0-9]+([0-9]+)?$ ]]
    then
        n[$count]=${arg}
        let "count += 1"
    else
        echo "$arg is not a valid integer!"
    fi
done

sort -n <(printf "%s\n" "${n[@]}")
  • The presented technique uses an array to store all integer values before sorting them.
  • All numeric values are given as command line arguments to the script.
  • The script tests whether each command line argument is a valid integer before adding it to the n array.
  • The sorting part is done using sort -n, which sorts the array numerically. If you want to deal with strings, then you should omit the -n option.
  • The printf command, after sort -n, prints every element of the array in a separate line whereas the < character tells sort -n to use the output of printf as input.

Run the sort script with the following command:

./sort.sh 100 a 1.1 1 2 3 -1

The output of sort.sh will resemble the following:

a is not a valid integer!
1.1 is not a valid integer!
-1
1
2
3
100

A Game Written in bash

This section will present a simple guessing game written in bash(1). The logic of the game is based on a random number generator that produces random numbers between 1 and 20 and expects from the user to guess them.

File: guess.sh
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
#!/bin/bash
NUMGUESS=0

echo "$0 - Guess a number between 1 and 20"

(( secret = RANDOM % 20 + 1 ))

while [[ guess -ne secret ]]
do
    (( NUMGUESS = NUMGUESS + 1 ))
    read -p "Enter guess: " guess

    if (( guess < $secret )); then
        echo "Try higher..."
    elif (( $guess > $secret )); then
        echo "Try lower..."
    fi
done

printf "Yes! You guessed it in $NUMGUESS guesses.\n"

Run the guess script:

./guess.sh

The output of guess.sh will resemble the following:

./guess.sh - Guess a number between 1 and 20
Enter guess: 1
Try higher...
Enter guess: 5
Try higher...
Enter guess: 7
Try lower...
Enter guess: 6
Yes! You guessed it in 4 guesses.

Calculating Letter Frequencies

The following bash script will calculate the number of times each letter appears on a file.

File: freqL.sh
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
#!/bin/bash

if [ -z "$1" ]; then
    echo "Usage: $0 filename."
    exit 1
fi

filename=$1

while read -n 1 c
do
    echo "$c"
done < "$filename" | grep '[[:alpha:]]' | sort | uniq -c | sort -nr
  • The script reads the input file character by character, prints each character, and processes the output using the grep, sort, and uniq commands to count the frequency of each character.
  • The [:alpha:] pattern used by grep(1) matches all alphabetic characters and is equivalent to A-Za-z.
  • If you also want to include numeric characters in the output, you should use [:alnum:] instead.
  • Additionally, if you want the output to be sorted alphabetically instead of numerically, you can execute freqL.sh and then process its output using the sort -k2,2 command.

Run the freqL script:

./freqL.sh text.txt

The output of freqL.sh will resemble the following:

   2 b
   1 s
   1 n
   1 i
   1 h
   1 a
Note
The file text.txt will not exist by default. You can use a pre-existing text file to test this script, or you can create the text.txt file using a text editor of your choice.

Timing Out read Operations

The read builtin command supports the -t timeout option that allows you to time out a read operation after a given time, which can be very convenient when you are expecting user input that takes too long. The technique is illustrated in timeOut.sh.

File: timeOut.sh
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
#!/bin/bash

if [[ $# -le 0 ]]
then
    printf "Not enough arguments!\n"
    exit
fi

TIMEOUT=$1
VARIABLE=0

while :
do
  ((VARIABLE = VARIABLE + 1))
  read -t $TIMEOUT -p "Do you want to Quit(Y/N): "
  if [ $VARIABLE -gt $TIMEOUT ]; then
    echo "Timing out - user response took too long!"
    break
  fi

  case $REPLY in
  [yY]*)
    echo "Quitting!"
    break
    ;;
  [nN]*)
    echo "Do not quit!"
    ;;
  *) echo "Please choose Y or N!"
     ;;
  esac
done
  • The timeout of the read operation is given as a command line argument to the script, an integer representing the number of seconds that will pass before the script will “time out” and exit.
  • The case block is what handles the available options.
  • Notice that what you are going to do in each case is up to you – the presented code uses simple commands to illustrate the technique.

Run the timeOut script:

./timeOut.sh 10

The output of timeOut.sh will resemble the following:

Do you want to Quit(Y/N): Please choose Y or N!
Do you want to Quit(Y/N): Y
Quitting!

Alternatively, you can wait the full ten seconds for your script to time out:

Do you want to Quit(Y/N):
Timing out - user response took too long!

Converting tabs to spaces

The presented utility, which is named t2s.sh, will read a text file and convert each tab to the specified number of space characters. Notice that the presented script replaces each tab character with 4 spaces but you can change that value in the code or even get it as command line argument.

File: tabs2spaces.sh
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
#!/bin/bash

for f in "$@"
do
    if [ ! -f $f ]
    then
      echo $f does not exist!
      continue
    fi
    echo "Converting $f.";
    newFile=$(expand -t 4 "$f");
    echo "$newFile" > "$f";
done
  • The script uses the expand(1) utility that does the job of converting tabs to spaces for us.
  • expand(1) writes its results to standard output – the script saves that output and replaces the current file with the new output, which means that the original file will change.
  • Although tabs2spaces.sh does not use any fancy techniques or code, it does the job pretty well.

Run the tabs2spaces script:

./tabs2spaces.sh textfile.txt

The output of tabs2spaces.sh will resemble the following:

Converting textfile.txt.
Note
The file textfile.txt will not exist by default. You can use a pre-existing text file to test this script, or you can create the textfile.txt file using a text editor of your choice.

Counting files

The following script will look into a predefined list of directories and count the number of files that exist in each directory and its subdirectories. If that number is above a threshold, then the script will generate a warning message.

File: ./countFiles.sh
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
#!/bin/bash

DIRECTORIES="/bin:/home/mtsouk/code:/srv/www/www.mtsoukalos.eu/logs:/notThere"

# Count the number of arguments passed in
if [[ $# -le 0 ]]
then
    echo "Using default value for COUNT!"
else
    if [[ $1 =~ ^-?[0-9]+([0-9]+)?$ ]]
    then
        COUNT=$1
    fi
fi

while read -d ':' dir; do
    if [ ! -d "$dir" ]
    then
        echo "**" Skipping $dir
        continue
    fi
    files=`find $dir -type f | wc -l`
    if [ $files -lt $COUNT ]
    then
        echo "Everything is fine in $dir: $files"
    else
        echo "WARNING: Large number of files in $dir: $files!"
    fi
done <<< "$DIRECTORIES:"

The counting of the files is done with the find $dir -type f | wc -l command. You can read more about the find command in our guide.

Run the countFiles script:

./countFiles.sh 100

The output of countFiles.sh will resemble the following:

WARNING: Large number of files in /bin: 118!
Everything is fine in /home/mtsouk/code: 81
WARNING: Large number of files in /srv/www/www.mtsoukalos.eu/logs: 106!
** Skipping /notThere

Summary

The bash scripting language is a powerful programming language that can save you time and energy when applied effectively. If you have a lot of useful bash scripts, then you can automate things by creating cron jobs that execute your bash scripts. It is up to the developer to decide whether they prefer to use bash or a different scripting language such as perl, ruby, or python.

More Information

You may wish to consult the following resources for additional information on this topic. While these are provided in the hope that they will be useful, please note that we cannot vouch for the accuracy or timeliness of externally hosted materials.

This page was originally published on


Your Feedback Is Important

Let us know if this guide was helpful to you.


Join the conversation.
Read other comments or post your own below. Comments must be respectful, constructive, and relevant to the topic of the guide. Do not post external links or advertisements. Before posting, consider if your comment would be better addressed by contacting our Support team or asking on our Community Site.
The Disqus commenting system for Linode Docs requires the acceptance of Functional Cookies, which allow us to analyze site usage so we can measure and improve performance. To view and create comments for this article, please update your Cookie Preferences on this website and refresh this web page. Please note: You must have JavaScript enabled in your browser.