Unix CLI: The Essentials

This article is adapted from the notes of the Unix@Home Workshop held in August 2020. It contains the essential knowledge that a CS2030S students need to survive using the PE nodes to do the labs (and to use a Unix CLI in general).

Learning Objectives

Students should

• be able to navigate comfortably in a Unix file system using shell commands cd, pwd, and ls.
• be able to use ., .., ~ or / to form a relative or absolute path name to a file.
• be able to organize files in a Unix file system using shell commands mkdir, rmdir, mv, rm, and cp.
• be able to view the content of the file using cat, more or less.
• be able to manage the permission of a file using chmod.
• be able to perform standard input/output redirection with < and >.

Unix Directory

Unix organizes files according to a directory structure. The following shows an example.

This structure is also known as a directory tree. There are two important directories that you need to know.

Root Directory

The root directory is the directory at the top of the directory tree¹. It is simply referred to as /, without any name. Under the root directory are many other system directory, which a casual user does not normally need to (and have no permission to) modify.

Home Directory

Each user has his/her own home directory. The above figure shows where the home directory of the user ooiwt resides in the directory tree. The user ooiwt may create files or directories in his/her home directory, but not elsewhere unless permission is given.

The home directory is referred to with the symbol ~ in bash. Sometimes we add the username behind ~ to indicate the home directory of the other user. E.g., ~bob means the home directory of a user named bob.

Current Working Directory

A user can navigate around the directory tree. The current working directory is the directory that the user is currently in. In contrast to the root and home directory, which are fixed², the current working directory changes as the user moves around. Knowing the current working directory is important since this is the default location in the directory tree a command executes. As such, many systems by default display the current working directory as part of the bash command prompt.

The current working directory is referred to with the symbol . in bash.

Parent Directory

The parent directory is directory one layer up from the current directory.

The parent directory is referred to with the symbol .. in bash.

To summarize, here are the short form representations:

Symbol	Meaning
/	the root directory
~	the home directory
.	the current working directory
..	the parent directory

Specifying a Path

To specify a directory or a file in the Unix directory tree, we can use either the absolute path or the relative path.

Absolute path

The absolute path is constructed as follows, starting from the root of the directory structure, find a path (a sequence of directories) to the location that you want to specify, then concatenate the names of the directories together, separated by the forward-slash /. This is a similar notation used for Website URLs, so you should already be familiar with it. For instance, the path /home/o/ooiwt is the absolute path of the directory named ooiwt in the figure above.

An absolute path is independent of the current working directory and always start with / or ~

Relative path

The relative path is dependent on the current working directory. To refer to another location, start from the current directory, and find a path (a sequence of directories) to the location that you refer to. When we go up a tree, we use .. to represent the directory.

For example, referring to the figure above, if we are in the directory /home/b, and we wish to refer to /home/o/ooiwt, we can use the relative path ../o/ooiwt. If we wish to refer to /home/b/bob, we can use the relative path bob.

A relative path never starts with /.

Directory-related Commands

Now, let's take a look at some basic commands available in bash that deals with navigation and directories.

pwd: Print Current Working directory

pwd shows you which directory you are currently in. Type pwd into the command prompt, and it will print the absolute path to your current working directory. For instance, Suppose you are in /home/o/ooiwt, entering

pwd

will give the output

/home/o/ooiwt

ls: LiSt content of a directory

The ls list the content in the current working directory.

Rule of Silence

Unix follows the economical rule of silence: programs should not print unnecessary output, to allow other programs and users to easily parse the output from one program. So, if ls has nothing to list, it will list nothing (as opposed to, say, printing "This is an empty directory.")

In Unix, a file or directory with a name that starts with . is hidden from ls. Common hidden files are .vimrc, the configuration file for vim, and .ssh, the directory to store your SSH keys and configuration. To view all the files, including hidden files, run ls with the -a flag (a stands for all).

mkdir: MaKe a subDIRectory

The mkdir command creates a subdirectory with the given name in the current directory.

In the example below, we assume that we start with an empty directory.

$ ls
$ mkdir workshop
$ ls
workshop
$ ls -F
workshop/

Here, we create a directory called workshop. Now, when we ls, you can see the directory listed.

You may also use ls -F for more information (-F is one of the many _options/flags available for the ls command. To see a complete list of the options, refer to the man pages, i.e., man ls_).

The slash / beside the filename tells you that the file is a directory. A normal file does not have a slash beside its name when "ls -F" is used.

You may also use the ls -l command (hyphen el, not hyphen one) to display almost all the file information, include the size of the file and the date of modification.

Use Up Arrow for Command History

bash maintains a history of your previously executed commands, and you may use the Ctrl+P (previous) and Ctrl+N (next) to go through it. Press the Ctrl+P until you find a previously executed command. You may then press Enter to execute it or edit the command before executing it. This is handy when you need to repeatedly execute a long bash command.

cd: Change Directory

To navigate in the directory tree, changing the current working directory from one to another, we use the cd command.

$ pwd
/home/o/ooiwt
$ cd workshop
$ pwd
/home/o/ooiwt/workshop

Suppose our starting working directory is /home/o/ooiwt, after we cd into workshop, the current working directory becomes /home/o/ooiwt/workshop. Note that cd can take in either an absolute path or a relative path. The example above takes in a relative path as the argument.

As mentioned in "Unix CLI: Background", it is common to include the current working directory into the shell's prompt. So, you may see your command prompt updated to include the new working directory.

Entering cd alone (without argument) brings you back to your home directory.

rmdir: ReMove a subDIRectory

rmdir removes a subDIRectory in the current directory -- note that a directory must be empty before it can be removed.

The command

$ rmdir workshop

will remove the directory that you just created.

File Management

This note assumes that you are familiar with navigation with the Unix directory structure.

We will now learn some commands to help us deal with files.

Let's recreate the workshop directory in case you have already deleted it with the rmdir at the end of the last unit.

$ mkdir -p workshop
$ cd workshop
$ ls

All the above commands should complete successfully and silently. Let's populate the directory with a new file. Cut-and-paste the command below into the command prompt:

wget https://www.comp.nus.edu.sg/~ooiwt/unix-workshop/test.txt

You should see a file being downloaded and saved with an output similar to below:

--2020-07-27 15:26:49--  https://raw.githubusercontent.com/nus-unix-workshop/2021-s1/master/test.txt
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.0.133, 151.101.64.133, 151.101.128.133, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.0.133|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 64 [text/plain]
Saving to: ‘test.txt’

test.txt              100%[======================>]      64  --.-KB/s    in 0s

2020-07-27 15:26:50 (2.35 MB/s) - ‘test.txt’ saved [64/64]

Now run ls, and you should see that ls returns test.txt as the content of the current working directory.

$ ls
test.txt

cp: CoPy files

Now let's try to copy this file to another name.

$ cp test.txt foo.txt
$ ls
test.txt foo.txt

The command above copies the file test.txt into foo.txt.

If you want to copy the whole directory, use -r flag, where r stands for copying recursively.

Now let's create another directory called copy.

$ cd ..
$ mkdir copy
$ cd copy
$ ls

Run pwd to double-check that you are in the directory called copy that is at the same level as workshop.

Now, we are going to use cp with the -r flag, to copy recursively the whole of workshop directory over.

$ cp -r ../workshop .

The command cp takes in two arguments, the first is the source, and the second is the destination.

Note that we use . above to indicate that we wish to copy the whole subtree of workshop over the current directory. The command should complete without any message. Upon completion, you can run ls to double-check that the workshop directory exists under workshop.

cp Overwrites

If there is an existing file with the same name, cp will overwrite the existing file without warning.

mv: MoVe or rename files

Now, let's change directory back to workshop.

$ cd ../workshop

and use the mv command to rename foo.txt into bar.txt.

$ ls
foo.txt test.txt
$ mv foo.txt bar.txt
$ ls
bar.txt test.txt

As you can see above, just like cp, mv takes in two arguments, the first is the source and the second is the destination.

If the destination of mv is a directory, however, instead of renaming, the mv commands move the source to the destination directory.

$ ls
bar.txt test.txt
$ mv ../copy/workshop/foo.txt .
$ ls
bar.txt foo.txt test.txt

Here, you can see that we have moved foo.txt over to the current directory.

mv Overwrites

If there is an existing file with the same name, mv will overwrite the existing file without warning. mv comes with a -i flag that interactively asks you if you are sure if you want to overwrite a file. It is a good idea to always run mv -i. Hit Y to continue overwriting the existing file. You can look up on the Internet on how to make -i the default using alias command in .bashrc.

Use Tab for Name Completion

If you have a very long file name, you may use the bash auto-completion feature to reduce typing. For instance, you may type:

$ mv t

and press the Tab key, bash will complete the filename for you if there is only one filename with the prefix "t". Otherwise, it will fill up the filename to the point where you need to type in more characters for disambiguation. The Tab key can also complete the name of a command.

rm: ReMove files

We can use rm to remove files. Be careful with this command -- files deleted cannot be restored. There is no trash or recycled bin like in Mac or Windows.

$ ls
bar.txt foo.txt test.txt
$ rm foo.txt
$ ls
bar.txt test.txt

rm -rf

While the Unix command line provides lots of flexibility and power, with great power comes great responsibility. Some commands are extremely dangerous. rm -rf * is the most famous one. The notation * refers to all files, and the flag -f means forceful deletion (no question asked!), and -r means remove recursively everything under the current directory tree. Accidentally running this command has ruined many lives. Read more here

rm comes with a -i flag that interactively asks you if you are sure if you want to delete a file. It is a good idea to always run rm -i.

$ rm -i bar.txt
rm: remove regular file 'bar.txt'?

Type y or n to answer yes or no respectively.

cat: CATenate file content to screen

To quickly take a look at the content of the file, use the cat command.

$ cat test.txt
This is a test file for learning Unix file management commands.

less is a variant of cat that includes features to read each page leisurely and is useful for long files.

$ less test.txt

In less, use <space> to move down one page, b to move Back up one page, and q to Quit.

man: Online MANual

An online help facility is available in Unix via the man command (man stands for MANual). To look for more information about any Unix command, for example, ls, type man ls. Type man man and refer to Man Pages to find out more about the facility. To exit man, press q.

File Permission Management

File permissions determine who can do what to a file. Typically, you do not need to fiddle with the file permission, but when you need to, it is usually for an important reason, and it is critical to do it right.

The What of File Permissions

Let's look at what you can do to a file first. Unix file permissions allow control of three actions: r (read), w (write) and x (execute). These permission settings allow the following different actions to be done for regular files and directories.

permission	effect on file	effect on directory
r	reading the content of a file	read the names of the files in the directory
w	writing into a file	create/delete/rename files in the directory
x	executing a file	access contents and meta-info (size, creation time) of files in the directory

These three actions can be controlled independently.

The permissions on a file can be expressed in two ways:

• using symbolic notation. For instance, rwx, r-x, -wx, where a - means that the corresponding permission is not given (in the order of r, w, x).

• using a numerical notation. This notation uses a digit between 0 and 7, which is computed as a sum of the individual digit representing the permissions: r is represented with 4, w is represented with 2, and x is represented with 1. For instance, r-x has a numerical representation of 5, and -wx has a numerical representation of 3.

The Who of File Permissions

Unix divides the users into three classes: u is the user who owns the file; g refers to the users in the same group as the user; and o are all the other users.

The permissions can be controlled separately for these classes of users. The permission notation simply concatenates the file permissions of each class of users together, in the order of u, g, and o.

For instance, the permission of 644, or rw-r--r--, on a file means that:

• the owner can read and write
• the group users can only read
• all the other users can only read

Checking file permission

You can view the permission of a file by using the ls -l command (l for long format):

$ ls -l test.txt
-rw-r--r--@ 1 ooiwt  staff  64 Jul 28 09:52 test.txt

Ignoring the first - and the last @, you can see that the permission of test.txt is 644.

The chmod command

You can use chmod command to change the permissions of a file or a directory.

For instance,

$ chmod 666 test.txt
$ ls -l test.txt
-rw-rw-rw-@ 1 ooiwt  staff  64 Jul 28 09:52 test.txt

would change add the permission w to both group and other users¹.

An alternative way is to just specify the changes. To remove the permission to write from others, you can run:

$ chmod o-w test.txt
$ ls -l test.txt
-rw-rw-r--@ 1 ooiwt  staff  64 Jul 28 09:52 test.txt

Common Scenarios for chmod

Here are some scenarios where you might need to use the chmod command:

• If you use the SoC Unix server to do your homework, you should prevent the directory that stores your homework from being accessible by other users. Make sure that your homework directory as the permission of 700.

• If you download a file from the Internet and you do not have the permission to read it, you should do a u+r to give yourself the read permission.

• A program should have execution permission to run. If you have a script or an executable file that you can't run, give yourself the execution permission u+x.

Standard Input/Output

History

Two of the design decisions of Unix that lead to its simplicity are the decisions to (i) decouple the physical input/output devices from the programs, allowing programs written for Unix to read from abstract input and output devices; and (ii) make all programs read and write from these abstract input and output devices by default. Before Unix, the older operating systems often require programmers to painstakingly set up connections to the teletype machines and other devices for reading inputs and printing outputs. With Unix, programmers can now focus on solving the tasks at hand and let Unix takes care of the input and output.

The two abstract devices that Unix provides are called standard input and standard output. By default, the standard input refers to the keyboard and the standard output is the terminal.

Examples using cat and wc

Let's look at these concepts closer, by examining some examples.

Remember cat? The cat command takes in a filename and it prints the content of the file to the standard output.

$ cat test.txt
This is a test file for learning Unix file management commands.

If no filename is given, cat by default try to read from the standard input. Try running:

$ cat

You will see that the command is waiting for you to type in something. Type in anything, as soon as you press Enter, cat is going to read in the text from the standard input, as if it is the content of a file, and then prints the content to the standard output. You can keep typing, supplying text to cat, or you can type Ctrl+D to send the end-of-input command to cat.

Let's look at another command, wc. wc is a utility that counts the number of lines, words, characters. If we call wc and supply it a file name, it will count the number of lines, words, and characters in that given file.

$ wc test.txt
       1      11      64 test.txt

The output means that there is 1 line, 11 words, and 64 characters in the file test.txt.

But if you do not pass in any file name, wc is going to read in the text from the standard input, as if it is the content of a file, and prints the three counters to the standard output. Go ahead and try:

$ wc

You will see that the wc command is waiting for you to type in something. Type in a few sentences, you can hit Enter for a new line. When you are done, type Ctrl+D. wc will count the number of lines, words, and characters for the text that you just entered.

Output Redirection

By defining two abstract input and output devices (or channels), Unix frees the programmers from worrying about where to read the input from and write the output to. Most of the time, we can write the output of the program to the standard output. In instances where we need to write the output to another location, we can just redirect the output.

The operators > and >> are used to redirect the standard output to a file. The difference is that > will overwrite the given file, while >> will concatenate into the given file.

For example:

$ wc test.txt > test.count
$ cat test.count
       1      11      64 test.txt

The first command redirects the output from wc to a file named test.count, so you do not see anything printed to the output anymore. We can check by running cat on the new file test.count -- indeed the original output from wc is now stored in the file test.count.

If we repeat the command wc test.txt > test.count again, you can see that the file has been overwritten with the output from wc again. But if we replace > with >>, a new line is concatenated into test.count. So the file now has two lines.

$ wc test.txt > test.count
$ cat test.count
       1      11      64 test.txt
$ wc test.txt >> test.count
$ cat test.count
       1      11      64 test.txt
       1      11      64 test.txt

Input Redirection

The operator < is used to redirect a file into the standard input. So, instead of reading from the keyboard, we can now read from a file. Commands such as cat and wc already support from a file directly, so there is no difference in terms of functionality to run the commands by passing in the file name, or by using the < operator.

$ wc test.txt
       1      11      64 test.txt
$ wc < test.txt
       1      11      64
$ cat test.txt
This is a test file for learning Unix file management commands.
$ cat < test.txt
This is a test file for learning Unix file management commands.

Note the slight difference in the output format of the second wc above -- it no longer prints the file name since from wc points of view, it is read from the standard input and not from a file, so it is not aware of the file named test.txt

In most CS programming assignments, however, to keep things simple, you will be asked to read from the standard input only, so the < is a great time-saver -- you do not have to repeatedly type in the same input data over and over from the keyboard. You can just save the input data in a file, then redirect it to standard input with the < operator.

Other Useful Online Tools

• explainshell explain each shell command for you

---

1. Giving write permission to other users is a security risk. You should not do this unless you know what you are doing. ↩↩

2. Not exactly true -- since Unix is designed to be flexible, even the root and the home directory can be changed! But let's not worry about that for now since there is no good reason to do that as a beginner. ↩