Skip to content

Download C03-Advanced-EN.pdf
Download SLIDE-C03-Advanced-EN.pdf

Math

Math in Bash base

Arithmetic operations in bash can only be performed on integers.

$(( INT OPERATION_TYPE INT))

# Ex 1: simple addition
username@hostname:~$ echo $(( 2 + 2 ))
4
username@hostname:~$ foo=$(( 2 + 2 ))
username@hostname:~$ echo $foo
4
username@hostname:~$ echo $(( $foo + 2 ))
6

Math in Bash operation type

The operations type are :

  • addition : +

  • substration : -

  • multiplication : *

  • division : /

Round in bash math

Bash rounds the result of operations "down" to get an integer.

username@hostname:~$ echo $(( 9 / 3 ))
3
username@hostname:~$ echo $(( 10 / 3 ))
3

Bc

If you need to perform arithmetic operation with float you need bc with its scale option :

The syntax is as follows:

username@hostname:~$ echo "scale=3;10/3" | bc
3.333
username@hostname:~$ echo "scale=2;10/3" | bc
3.33
username@hostname:~$ echo "scale=1;10/3" | bc
3.3

Sequences

Sometimes it's usefull to generate a sequence of integer from a start to an end point. You can use the seq command to do so.

Example :

username@hostname:~$ seq 1 10
1
2
3
4
5
6
7
8
9
10

The flows

Each process has 3 flows :

  • The input stream "stdin"
  • The standard output stream "stdout"
  • The error output stream "stderr"

Stdin

This is an invisible but very useful feed. It's an input data of a command, generally from the keyboard or from another command.

Stdin (example)

Read from the keyboard gradually:

username@hostname:~$ sort << ENDSORT
> toto
> titi
> tata
> ENDSORT
tata
titi
toto

Terminal wait for some text. As long as the character string != keyword (here ENDSORT), continue. All data sent to sort until ENDSORT keyword detected. The result sort will be in its stdout. We can modify the stdout :

username@hostname:~$ sort << ENDSORT > file_sort.txt

Stdout

This is the most visible flow. By default, it will be displayed on the screen.

It is possible to :

  • send it in a new file : cmd > file
  • send it at the end of a file : cmd >> file
  • send it to another command to make a chaining commands : cmd1 | cmd2 (the stdout of cmd1 will be the stdin of cmd2)

Flux Linux

Stdout example (1/2)

username@hostname:~$ echo "this is my stdout"
this is my stdout

Stdout example (2/2)

username@hostname:~$ echo -e "toto\ntata" >> File.txt
username@hostname:~$ cat File.txt
toto
tata
username@hostname:~$ cat File.txt | grep "to" 
toto
username@hostname:~$ cat File.txt | grep "to" | wc -l
1

Stderr

This is the flow that the terminal displays during an error. For this flow it is possible to :

  • send it in a new file : cmd 2> file
  • send it at the end of a file : cmd 2>> file
  • send it to stdout : cmd 2>&1 > file or cmd &> file (the set of stderr and stdout will be in file)

Stderr (example)

username@hostname:~$ ls 
File
Directory
username@hostname:~$ cat FileNotExist
cat: FileNotExist: No such file or directory
username@hostname:~$ cat FileNotExist 2> ./stderr.txt
username@hostname:~$ cat FileNotExist 1> ./stdout.txt
username@hostname:~$ cat stderr.txt
cat: FileNotExist: No such file or directory
username@hostname:~$ cat stdout.txt

username@hostname:~$

ReGex

The different types

ReGex allow you to match specifics patterns (like a phone number, an email address ...)

There are 2 types of regular expressions :

  • Basic regular expressions (vi, grep, expr, sed) : ERb
  • Extended regular expressions (grep -E, egrep, awk) : ERe

The keywords of the regex

  • "^" = Start of line : "^toto"
  • "\$" = End of line : "toto\$"
  • "." = Any character
  • "?" = 0 or 1 time the previous character or grouping : "?t"
  • "*" = 0 to n times the previous character or grouping : "*t"
  • "+" = 1 to n times the previous character or grouping : "+t"
  • "\" = Protection of a special character
  • "[list_of_chars]" = A quoted character in the list : "[a-z]" "[A-Z]" "[0-9]"
  • "[^list_of_chars]" = A character that is not mentioned in the list : "[^0-9]"
  • "(PATERN){n}" = Number of times of patern (ERe) : "(ti){2}"

Example of ReGex (1/3)

Let's take this text file as an example :

toto likes titi
Isen 2021
Toto likes titi

  • Find lines that start with toto and end with titi
isen@isen : cat exampleRegex.txt
toto likes titi    
Isen 2021    
Toto likes titi  
isen@isen : grep "^toto.*titi$" exampleRegex.txt
toto likes titi

Example of ReGex (2/3)

  • Find lines that start with a capital letter 
isen@isen : cat exampleRegex.txt
toto likes titi    
Isen 2021    
Toto likes titi 
isen@isen : grep "^[A-Z]" exampleRegex.txt
Isen 2021
Toto likes titi

Example of ReGex (3/3)

  • Find lines that contain "to" 2 times 
isen@isen : cat exampleRegex.txt
toto likes titi    
Isen 2021    
Toto likes titi
isen@isen : grep -E "(to){2}" exampleRegex.txt
toto likes titi

Grep

Introduction

Grep command can be used to find or search a regular expression or a string in a text file.

Searching in a specific file

Searching the pattern "Toto" in the file exempleRegex.txt.

isen@isen : grep "Toto" exempleRegex.txt
Toto love titi    
Totope

Searching recursivly

Searching the pattern "Toto" recursivly in all files in /home/isen directory and sub directory.

isen@isen : grep -r "Toto" /home/isen
Toto love titi    
Totope    
Toto hate titi

Ignore case sensitivity

Searching the pattern "toto" by ignoring case sensitivity in the file exempleRegex.txt.

isen@isen : grep -i "Toto" exempleRegex.txt
Toto love titi    
Totope    
toto tata
TOTo titi

Sed

Introduction

Sed is a command that allows manipulation of text file from regular expression.

By default sed displays the result in the stdout. To do the action directly in the file you need the option -i.

Substitution

Change one patern by another.

sed "s/PATERN_TO_LOOK_FOR/REPLACEMENT_PATTERN/"

isen@isen : cat exampleRegex.txt
toto likes titi    
Isen 2021    
Toto likes titi 
isen@isen : sed "s/titi/loic/" exampleRegex.txt
toto likes loic    
Isen 2021    
Toto likes loic

Insert

Add a line before the wanted patern.

sed "/PATERN_TO_LOOK_FOR/iPATERN_TO_ADD/"

isen@isen : cat exampleRegex.txt
toto likes titi    
Isen 2021    
Toto likes titi
isen@isen :  sed "/Toto/iTiti" exampleRegex.txt
toto likes titi    
Isen 2021 
Titi   
Toto likes titi

Append

Add a line after the wanted patern.

sed "/PATERN_TO_LOOK_FOR/aPATERN_TO_ADD/"

isen@isen : cat exampleRegex.txt
toto likes titi    
Isen 2021    
Toto likes titi 
isen@isen :  sed "/toto/aTiti" exampleRegex.txt
toto likes titi   
Titi 
Isen 2021    
Toto likes titi

Delete

Delete a line containing a pattern (by modifying permanently the file with -i option).

sed "/PATERN/d"

isen@isen : cat exampleRegex.txt
toto likes titi    
Isen 2021    
Toto likes titi
isen@isen : sed -i "/titi$/d" exampleRegex.txt    
isen@isen : cat exampleRegex.txt 
Isen 2021