14 KiB
Playing with text on the command line
The command line (also known as the command line interface, or CLI, or sometimes the terminal), is a plain text-based interface for executing commands on a computer. If you've ever seen a movie about hackers from the 1980s, like WarGames, where they stare at a prompt on a black screen and type in commands one at a time, it's basically that.
You have a prompt, and you can type in a command and hit 'Enter' to execute it. An example command would be:
touch newfile.txt
This command will create a file called newfile.txt.
How to access the command line
Mac OS X: Go to /Applications/Utilities and click on "Terminal" or search for "Terminal" in Spotlight.
Desktop Linux: You can search for the "Terminal" application from the Dash. Let's be honest, though, if you're running Linux, you probably don't need this tutorial.
Windows: Windows is a bit of a special case. If you go to the Start Menu and click "Run", and then type "cmd" and hit enter, it will open the Windows version of the command line. Unfortunately, the Windows version of the command line kind of has its own system, so for the purposes of following these examples, you'll want to install Cygwin, which will allow you to mimic a Linux-style command line:
A little more detail
Commands generally take the format:
[name of the command] [option] [option] [option] ...
The prompt will also show what directory you're currently sitting in. Whenever you execute a command, you do it from a particular directory. This matters because when you execute a command that involves a filename or a directory name, you can specify it one of two ways:
Relative Paths
Specifying a file or directory as a relative path means you are specifying where it sits relative to the directory you're in. For example, let's say you're in the videos subdirectory of the files directory. You'll see this prompt:
/files/videos$
If you execute a command like touch newfile.txt, it will create newfile.txt inside the current directory. Relative paths don't start with a slash.
Absolute Paths
Specifying a file or directory as an absolute path means you are specifying where it sits on the computer in absolute terms, starting from the top level. For example, let's say you're in the videos subdirectory of the files directory again.
/files/videos$
If you execute a command like touch /files/music/newfile.txt, it will create newfile.txt inside a different folder, the music subfolder of the files folder. Absolute paths start with a slash.
If you use an absolute path, the command will do the same thing no matter what directory you execute it from.
So these two commands will have the same result from the /files/videos directory:
/files/videos$ rm video.mp4
(This will delete the file `video.mp4` from the current directory)
/files/videos$ rm /files/videos/video.mp4
(This will delete `video.mp4` from the /files/videos/ directory, which happens to be the current directory)
The same two commands will not have the same result if you are in a different directory:
/files/text$ rm video.mp4
(This will try to delete the file video.mp4 from the 'text' subdirectory instead, because that's the current directory)
/files/text$ rm /files/videos/video.mp4
(This will delete the file from the /files/videos/ directory, even though it isn't the current directory)
Remember:
Starting a path with a slash means you want to give the entire path and ignore what directory you're currently in. Not starting a path with a slash means you want to give the path starting from the directory you're in.
If you're ever unsure of what directory you're in, you can use the pwd (Print Working Directory) command to get the absolute path of the current directory.
~$ pwd
/Users/Noah
File Patterns
In most cases when you have to specify a file name or directory name, you can also specify a general pattern that might match multiple files. There are lots of ins and outs with this, but the most basic version is using the asterisk (*), which matches anything. It's also known as a wildcard.
Delete any file in the current directory
/files$ rm *
Delete any file that ends in '.txt'
/files$ rm *.txt
Delete any file that starts with 'data'
/files$ rm data*
Navigating
The two core commands for navigating what directory the prompt is in are cd and ls.
cd is a command to change the current directory, and must be followed by a directory you want to change to. You can supply an absolute or relative path.
This will put you in /files/videos
/files$ cd videos
/files/videos$
This will put you in /videos, and then the vines subdirectory
/files$ cd /videos
/videos$ cd vines
/videos/vines$
You can jump multiple levels at once if you want.
This will put you in /files/videos/short
/files$ cd videos/short
You can use cd .. to move up one level to the parent directory.
This will put you in /files
/files/videos$ cd ..
ls will list the files in the current directory. It's helpful for figuring out where you are, what files exist, and what subfolders exist.
/photos$ ls
thumbnails photo1.jpg photo2.jpg
Using ls -l will print the list vertically, with lots of other extra information about the file size, permissions, and last modified date:
/photos$ ls -l
-rw-rw-r-- 1 noah noah 58133 Oct 22 17:13 photo1.jpg
-rw-rw-r-- 1 noah noah 75640 Oct 22 17:13 photo2.jpg
drwxrwxr-x 2 noah noah 4096 Oct 22 17:13 thumbnails
When typing in a directory or file name, you can hit the 'Tab' key to autocomplete if it's possible. For example, in the /photos folder, if you type in:
/photos$ cd thu
and hit 'Tab,' it will fill in the rest and show you:
/photos$ cd thumbnails
However, if there is more than possible file/directory that matches what you've typed so far, it won't work. If you type:
/photos$ rm pho
and hit 'Tab,' nothing will happen because you could be on your way to photo1.jpg OR photo2.jpg.
Command Output
The commands we're going to talk about all output their results as text. When you execute the command by hitting 'Enter', it will print out a bunch of output on extra lines below the prompt. For example, head [file] will print out the first 10 lines of a file.
/files$ head names.txt
Dan Sinker
Erika Owens
Noah Veltman
Annabel Church
Friedrich Lindenberg
Sonya Song
Mike Tigas
Brian Abelson
Manuel Aristaran
Stijn Debrouwere
/files$
Notice that after it prints out its output, it goes back to giving you a fresh prompt. Getting the output printed out to you in this fashion is useful if you're just poking around, but often you want to do one of two things: send the output to a file, or send the output to another command as an input.
Sending the output to a file
You can send the output to a new file this way:
/files$ head names.txt > first10names.txt
If first10names.txt doesn't exist, it will be created. If it already exists, it will be overwritten.
You can append the output to the end of an existing file this way:
/files$ head names.txt >> allnames.txt
This will add the output as 10 new lines at the end of allnames.txt.
Sending the output to another command as an input
You can send the output to another command using the pipe symbol (|). The grep command searches through some text for matches (more on this later), so you could do this to get the first 10 lines of a file, and then search for "Steve" within those 10 lines:
/files$ head names.txt | grep "Steve"
This is basically the same as doing this:
/files$ head names.txt > temporaryfile.txt
/files$ grep "Steve" temporaryfile.txt
But instead of first sending the output to a file and then running the second command on that file, you pipe the output directly from the first command into the second. You can chain as many of these together as you want:
/files$ grep "United States" addresses.csv | grep "California" | head
This would search the file addresses.csv for lines that contain the phrase "United States", then search the results for lines that contain the word "California", and then print out the first 10 of those matches.
Grep
The grep command will let you search a file (or multiple files) for a phrase. By default, it will print out each line that matches your search.
Print out lines that contain the word "darkwing":
/files$ grep "darkwing" famousducks.txt
Same as above, but the search is case-insensitive:
/files$ grep -i "darkwing" famousducks.txt
Find matches for the exact word "Donald" in a file - words that contain "Donald," like "McDonald," won't count:
grep -w "Donald" famousducks.txt
Find matches for "McDuck" in every file in the current directory:
grep "McDuck" *
Find matches for "McDuck" in every file in the current directory AND every subdirectory, all the way down:
grep -r "McDuck" *
For each match of "Howard", print out that line AND the 4 lines after it (5 lines total):
grep -A 4 "Howard" famousducks.txt
For each match of "Howard", print out that line AND the 4 lines before it (5 lines total):
grep -B 4 "Howard" famousducks.txt
For each match of "Howard", print out that line AND the 4 lines before it AND the 4 lines after it (9 lines total):
grep -C 4 "Howard" famousducks.txt
Instead of printing out the matching lines themselves, print out the filenames that match your search:
grep -l "Daffy" *
Just get the number of matches:
grep -c "Daffy" *
Show line numbers along with the matching lines:
grep -n "Daffy" famousducks.txt
Cat
The cat command will combine multiple files together. This will print three files in a row, as if they were one file:
cat turkey.txt duck.txt chicken.txt
Remember that this will just print the output into your terminal. More likely, you want to create a new file that combines them:
cat turkey.txt duck.txt chicken.txt > turducken.txt
turducken.txt will contain all of the lines in turkey.txt, followed by all of the lines in duck.txt, followed by all of the lines in chicken.txt.
If you want to combine ALL of the files in a directory, you can use a wildcard:
cat * > allfilescombined.txt
Head
The head command will print out the first 10 lines of a file:
/files$ head names.txt
You can also specify a different number of lines. This will print out the first 15 lines of a file:
/files$ head -n 15 names.txt
Or, if you want to print all the file but leave out the LAST 15 lines, you can give a negative number:
/files$ head -n -15 names.txt
One of the nice uses of head is to quickly peek inside a large text file to see what's in it without having to wait for a text editor to load it. This becomes a big deal when you're talking about a 1 GB file!
Tail
The tail command is the reverse of head. It will print out the last 10 lines of a file:
/files$ tail names.txt
This will print out the last 15 lines of a file:
/files$ tail -n 15 names.txt
Or, if you want to print all the file but leave out the FIRST 15 lines, you can add a plus sign:
/files$ tail -n +15 names.txt
This is helpful if you want to, say, remove a header row from a CSV file:
/files$ tail -n +1 names.txt > names-no-header.txt
Miscellaneous
If you just want to print out the entire contents of a file into your terminal, you can use cat and not combine it with anything. This is sort of against the whole point of cat, but is a handy trick.
/files$ cat address.txt
1600 Pennsylvania Avenue
Washington, DC 20500
If you want to get serious and open a file in a text editor that comes built in to your terminal, you can try nano:
/files$ nano address.txt
How many lines are in names.txt?
/files$ wc -l names.txt
18
Regular expressions
When using something like grep to search, you can search for a simple term with only letters, numbers, and spaces. But if you want to search for a pattern, you can use what's called a regular expression. Regular expressions use special characters to represent patterns, like "any number," "any letter," "X or Y," "at least three lowercase letters," and so on.
We won't worry about the ins and outs for now, but one useful operator is the period (.). In regular expression-ese, this means "One of any character." So you can search for something like:
/files$ grep -i "car.s" dictionary.txt
This would match words like cards,carts,cares, and so on. It would also match the middle of the phrase "scar story" (CAR S) because "any character" means ANY character, including a space or a punctuation mark.
One more example:
/files$ grep -i ".e.st" dictionary.txt
This would match things like least,beast, and heist.
More than one way to skin a cat
There are often lots of equally legitimate commands or combinations of commands to achieve the same purpose.
Example:
/files$ head -n 12 names.txt | tail -n 5
(Print out the first 12 lines, and then print out the last 5 lines of that)
is the same as
/files$ tail -n +7 names.txt | head -n 5
(Print out everything but the first 7 lines, then print the first 5 lines of that)
is pretty much the same as:
/files$ tail -n +7 names.txt > temporaryfile.txt
/files$ head -n 5 temporaryfile.txt
/files$ rm temporaryfile.txt
(Save everything but the first 7 lines to a temporary file, then print the first 5 lines of that, then delete the temporary file)
Questions/Comments/Suggestions
Original Repo modified by Frenchris for 01-edu learning purposes
Be aware that some or all the files may have been modified compared to the original.
Credits of original repo:
Noah Veltman Web: http://noahveltman.com Twitter: @veltman Email: noah@noahveltman.com