Bash Reference_12550194

---

title: "TheNessLab : Bash Reference"

1. TheNessLab
2. The Ness Lab Wiki
3. Coding tips and tutorials

TheNessLab : Bash Reference

Created by Ahmed Raza Hasan, last modified on Jun 01, 2017

(*** under construction - turn back now, and have your mortal life spared ***)

The bash shell is a highly useful means of interacting with your computer (or, in the case of our server, a computer). We all use the bash shell in order to get our Jupyter Notebooks set up and perhaps run some command line programs, but turns out it can do a lot more past that! 

Understanding even basic Unix commands can go a really long way, especially in a genomics environment where the inputs are big and the outputs can be bigger. Working with large text files in particular is a breeze with said tools where something like Python or R would struggle to even load them in. Finally, bash is incredibly powerful when it comes to automating tasks with ease.

A quick glossary:

• stdin = input. this is what you are entering into the command line
• stdout = output. this is what gets printed to the console after something is run

fundamentals

pwd - prints current directory to stdout

• the current directory you are in determines what bash 'immediately sees' when you, say, call upon a file
  • if this makes no sense, I promise it will soon enough

whoami - in case of identity crisis

ls - lists files in current directory

• super useful!
• like most Unix commands, can be altered with flags
  • flags are instantiated using a single hyphen (-), and almost always consist of just one letter
  • caveat - sometimes flags are entire words, and start with two dashes (ie --flag) but you likely won't see these as often
• for instance, ls -l will return files in the directory in a list format, while ls -t lists them in descending order since they were last modified
• multiple flags can be used at once, either as ls -l -t or just ls -lt

man - pull up the help page for a command

• did you forget the ls flags already?
• well, no sweat - running man ls will pull up a handy reference
• scroll up and down with the arrow keys, and press q to exit the man (manual) page

cd - change directory

• switches working directory to specified folder
• for instance, cd Desktop/ will take me to a folder called Desktop/, *assuming *Desktop/ is in my current directory
• obviously this isn't always the case - but to go back a folder, use cd .. or cd ../
  • this can be chained together using forward slashes - for instance, cd ../../ will go back two folders up in the hierarchy (i.e. from /scratch/research/projects to /scratch/)
• if the path given to cd starts at the root directory, then bash won't complain about being given a folder that's not in your current working directory
  • for instance, I can do cd /scratch/research/references even if I'm currently in scratch/research/projects/chlamydomonas, as long as I write out the full path

echo - prints input back to stdout

• bash's equivalent of print(), really
• echo hello there will return hello there
• inputs don't *have* to be in quotes, but bash won't mind if you do wrap them with either single or double quotes

basic file operations

cat - print contents of file

• cat <file> will simply return the contents of the file to stdout
• however, cat can take multiple files as arguments, and will *concatenate *them before printing to stdout
• for instance, cat <file1> <file2> will return the entire contents of both file1 and file2 one after the other
• this is useful for quickly combining large flat files containing data, and synergizes super well with pipe operations (more on that in the tips and tricks article)

mkdir - creates a new folder

• mkdir myfolder will create an empty folder called myfolder in your current working directory
• mkdir folder1 folder2 will make two folders named folder1 and folder2 respectively

mv - move files around/rename files

• mv file.txt data/files will move file.txt to /data/files
• adding the -v flag (i.e. mv -v) is especially useful, as bash will explicitly print out what it's moving and where
• mv works well with wildcards in moving several files around - more on that later on as well
• finally, mv also allows us to rename files by 'moving' a file on top of itself
  • for instance, mv file.txt my_file.txt will rename file.txt to my_file.txt

rm - delete things

• this is a frighteningly efficient command, and has no undo button
• in Unix, there is no Recycle Bin for deleted files to go to - they are promptly sent to the void and you will never, ever see them again. you hear that? never.
• rm -i will prompt for confirmation before you delete something, so it might be good to make a habit of that
• rm will complain if you make it delete a folder, but that can be bypassed with rm -rf
• possibly goes without saying, but be extremely careful when using rm on multiple files at once!

rmdir - delete a folder

• an alternative to the scarier rm -rf if you feel like it

slightly less basic file operations

chmod

head

tail

less

sort

wc

gzip

operators

pipe

>

>>

!$

!!

loops

parallel operations

vi 101

something about tmux?

Document generated by Confluence on May 22, 2024 11:44

Atlassian