This post list a few tips that I discovered when I started working back in 2013. I have improved my knowledge a great deal over the last years but the informations there can still be useful.

Bash (any)

Use a default value if the variable is unset or null (see more on parameter expansion):

echo "${TEST:-default}"

Transfer the parameters to a command:

test.sh:
  my_function "$@"

test.sh "this has spaces" 25
=> my_function will correctly receive 2 parameters: "this has spaces" and 25. If you use $@ without quotes or $*, the quoting would be lost and my_function would receive 4 parameters: "this", "has", "spaces" and "25"

Repeat last command :

!!

Substitute A for B in the last command and execute it:

^A^B^

Use the last item of the last command in the current one:

ls /tmp /tmp2
cd !$
pwd -> /tmp

For the first item of the last command (^ and $ match regexes):

ls /tmp /tmp2
cd !^
pwd -> /tmp2

Print every command before execution. Very useful when debugging bash scripts:

set -x

Exit the script at the first error (Note: beware grep)

set -e

Generally speaking, to increase the reliability of your scripts and prevent silent errors, this construct should be preferred:

set -eu
set -o pipefail

• -e: exit on error (warning: will not raise an error if the failing process is part of a pipe)
• -u: using undefined variables trigger an error
• -o pipefail: if a process fail in a pipe, trigger an error

Zsh

List any compressed file in any sub-directory **/* ending with either bz2 or gz with a size greater than 10MB (Lm+10)

ls -lh **/*.(bz2|gz)(Lm+10)

Print only the basename (t), the dirname (h) or the file name and directory without the extension (r) of the variable FILE

ls "${FILE:t}"
ls "${FILE:h}"
ls "${FILE:r}"

Arrays

Create a new Array:

IDs=(ber_sim boi_mar bom_cel bou_pat che_elo cha_chr)

Print length of Array:

echo "${#IDs}"

Loop

Print numbers from 1 to 5, instead of using seq:

for i in {1..5}; do echo $i; done

find

Print the directory (%h), the file name (%f) and the size in kB (%k) of any tar.bz2 archive bigger than 20MB in subdirectories:

find -iname "*.tar.bz2" -size +20M -printf '%ht%ft%kkn'

Copy any .sh file in subdirectories to the current directory:

find -name "*.sh" -exec cp '{}' . ;

awk

Select the second column of each line from command qjob, and for each selected part of line, execute qdel

qjob -l | awk '{print $2}' | while read i; do qdel $i

print the 4th value divided by 172 followed by a semicolon and the 8th value of each line of the file File.log . Two delimiters are used at the same time: : and ,

awk -F':|,' '{print $4/172 ";" $8}' File.log

More Infos

How to make portable tests in bash, or what are the differences between [ and [[