Scripting examples¶

Semicolon Usage¶

: is equivalent to true command. When used with another bash substitution command for example, the command return true instead of the string. Example:
${TEST1:=default1}
Assign default1 to TEST1 variable and return default1 (so, if the return string is a command, it will be interpreted by shell)

: ${TEST1:=default1}

Assign default1 to TEST1 variable and return true

Rename multiple files recursively¶

find . -iname <file_name_regex> -exec rename <old_name> <new_name> '{}' \;

A cool tool to use in one folder is rename tool

Rename multiple files in same folder¶

rename 's/<from>/<to>/' *.html

Replace in file and diff¶

sed -e 's/<from>/<to>/g' <file_name> | diff <file_name> -

Strict execution¶

set -euo pipefail

set -e : If command fails, make the script exit (not resuming on next line)
set -u : Treat unset variables as an error and exit
set -o pipefail : cause a pipeline (command | command) to fail if one of the subcommand fails
set -x : print each command before executing it

Parameter substitution¶

Default Value ${:-}¶

var=${parameter:-defaultValue}

Default Value if not set ${:=}¶

echo ${parameter:=defaultValue}

Important

This method do not work to assign values, use ${:-} instead.

Display message if $value is not passed ${:?}¶

var="${1:?Please provide var}"

Display message and run command ${:?}¶

_message="Please provide var"
var="${1:? $_message $(command)}"

Variable length ${#}¶

length=${#var}

Remove pattern - Front ${#} ${##}¶

url="https://www.google.com/test/mypage.html"

pagepath=${url#https://}    # Return "www.google.com/test/mypage.html"
pagename=${url#*/}      # Return "/www.google.com/test/mypage.html"
pagename=${url##*/}     # Return "mypage.html"

Remove pattern - Back ${%} ${%%}¶

url="https://www.google.com/test/mypage.html"

pagepath=${url%.html}       # Return "https://www.google.com/test/mypage"
pagename=${url%/*}      # Return "https://www.google.com/test"
pagename=${url%%/*}     # Return "https:/"

Find and replace ${//}¶

var="${x/unix/linux}"       # Replace first match
var="${x//unix/linux}"      # Replace all matches

Substring starting character ${::}¶

${parameter:offset:length}

x="nixcraft.com"
echo ${x:3:5}"          # Print craft

Work with arguments/positional parameters¶

Interesting values¶

Value	Meaning
$0	Name of script
$N/${N}	Nth argument passed to script (brackets are mandatory after $9)
$#	Number of arguments passed to script
$@	All arguments passed to script

Check if argument is an option¶

if [ "${1#-}" != "$1" ]; then
        ...
fi

Check if argument is some value¶

if [ "$1" = 'some-value' ]; then
        ...
fi

Add positional parameters before arguments¶

The following script adds newcommand before other positional parameters.

set -- newcommand "$@"
echo "$@"

Check if string is part of string¶

string='My long string'
if [[ $string == *"My long"* ]]; then
  echo "It's there!"
fi

For each argument¶

for var in "$@"
do
    echo "$var"
done

Obfuscate password¶

echo ${PASSWORD//[[:ascii:]]/*}

Print decorated messages¶

print_decorate_message () {
        message=$1
        message_length="${#message}"
        message_decorate_length=$((message_length + 2))
        decorate_line=$(printf "%-${message_decorate_length}s" "-")

        echo "+${decorate_line// /-}+"
        echo "| ${message} |"
        echo "+${decorate_line// /-}+"
}

Get script file path and name¶

SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
SCRIPT_NAME="$(basename "$0")"

If string is empty¶

if [ -z "$my_string" ]
then
        echo "String is empty"
fi

Insert line at specific line¶

line_number=3
new_line="my new line"
sed -i "${line_number}i${new_line}" my_file

Insert line after a specific line¶

line_number=3
new_line="my new line"
sed -i "${line_number}a${new_line}" my_file

Replace some text on line matching with word¶

sed -i '/<word>/s/<from>/<to>/' my_file

Grep exclusion¶

grep -r --color --exclude-dir={custom,lib,scripts} --exclude={*.xml,error_log} "beta" .

Extract string match between words from file¶

Grep way¶

grep -oP '(?<=<from>).*(?=<to>)' <my_file>

Sed way¶

sed -rn 's/<from>(.*)<to>/\1/p' <my_file>

Parse json from command line¶

Useful documentation Use jqcommand to parse output Example:

curl -X GET "<url>" | jq

output | jq ".[].project_id"

Get the sum of array with jq¶

output | jq -s add/length

Get max value of array with jq¶

output | jq max