The pipe (|) and tee commands in Linux and Unix (macOS)

Sunday, July 14, 2024 at 5:21 PM | 3 min read

Last modified on Sunday, July 14, 2024 at 5:21 PM

#command line, #linux, #unix, #macOS, #pipe, #tee, #wc

Tee

Photo by Kampus Production on pexels.com

I just finished a article entitled Using pipe (|) and grep in Linux and Unix (macOS). In this article, I will discuss using pipe (|) and tee in Linux.

As discussed in my article Using pipe (|) and grep in Linux and Unix (macOS), the pipe (|) command, aka un-named pipe, transmits the output of one command to another command. It can redirect stdout, stdin, or stderr of one process to another for further processing.

The tee command reads the stdin and writes it to both stdout and one or more files. According to Geeks for Geeks,

The (tee) command is named after the T-splitter used in plumbing.

Geeks for Geeks also states,

It basically breaks the output of a program so that it can be both displayed and saved in a file. It does both the tasks simultaneously...

The best way to describe what tee does is by example. Let's say I ran the following in Terminal from inside a folder called text-files:

ls | tee tee.txt

It would return the following in Terminal:

duplicates.txt
first-named-pipe
list-home-directory.txt
names1.txt
names2.txt
tee.txt

And if I ran cat tee.txt, the following would be returned in Terminal:

duplicates.txt
first-named-pipe
list-home-directory.txt
names1.txt
names2.txt
tee.txt

The stdout of the ls command is "piped" into the tee command via | as stdin, and then is both displayed as stdout in Terminal and written in tee.txt as stdout.

And if I ran the following command:

history | tee history.txt

Something like the following is output in Terminal:

  1 cd ..
  2 cd Desktop
  3 ls
  4 touch file1.txt
  5 vim file1.txt
  6 sudo apt install vim
  7 vim file1.txt
  8 expand file1.txt
  9 expand file1.txt > file2.txt
 10 cat file2.txt
 11 unexpand -a file2.txt
 12 mv file1.txt names1.txt

...

372 ls | tee -a names1.txt names2.txt tee.txt
373 cat tee.txt
374 ls | tee tee.txt
375 ls
376 cat tee.txt
377 history | tee history.txt

And the same content as output to Terminal is saved to a newly created file called history.txt, thanks to the tee command.

And if I run cat history.txt | tee history.txt history2.txt in Terminal, the exact same content as stdout to Terminal and history.txt is saved to a newly created history2.txt file which did not previously exist, thanks to the tee command.

  1 cd ..
  2 cd Desktop
  3 ls
  4 touch file1.txt
  5 vim file1.txt
  6 sudo apt install vim
  7 vim file1.txt
  8 expand file1.txt
  9 expand file1.txt > file2.txt
 10 cat file2.txt
 11 unexpand -a file2.txt
 12 mv file1.txt names1.txt

...

372 ls | tee -a names1.txt names2.txt tee.txt
373 cat tee.txt
374 ls | tee tee.txt
375 ls
376 cat tee.txt
377 history | tee history.txt

Let's say I ran more commands, and then wanted to append these new commands added to Terminal history to the bottom of the history.txt file. Running history | tee history.txt would erase all previous content and replace it with the new history. In order to make sure that the new commands are appended to the end of the file and don't overwrite its previous contents, I run the following in Terminal:

history | tee -a history.txt

This results in the following in Terminal:

...

378 ls
379 cat history.txt
380 cat history.txt
381 cat history.txt
382 history | tee history.txt

And inside history.txt:

...

378 ls
379 cat history.txt
380 cat history.txt
381 cat history.txt
382 history | tee history.txt

The -a flag, aka --append, is analogous to the output redirector append operator (>>). It appends new data to the end of a file, thereby not overwriting what is already there.

Let's say I run the following command in Terminal:

cat history.txt | tee history.txt history2.txt

This command returns the complete contents of history.txt in Terminal and copies its content into a newly created history2.txt. And if I run this command again after a bit of time and a few more commands, and add the -a or --append flag:

cat history.txt | tee -a history.txt history2.txt

And then I run the following command:

wc -l history2.txt

it returns the following in Terminal:

804 history2.txt

And when I ran:

wc -l history.txt

The wc command finds the number of lines, words, characters, bytes, and maximum line length of files, in that order.

the following is returned in Terminal:

804 history.txt

When I ran the cat command again, the contents of history.txt were appended to its previous content, thereby doubling the number of lines in the file. The same went for history2.txt. its number of files doubled as well.

If I run the following command:

wc history.txt

the following is returned:

804 2094 21800 history.txt

804 represents the number of (new) lines, 2094 represents the number of words, adn 21800 the number of bytes. And if I ran the following in Terminal:

wc -L history.txt

it returns the following:

67 history.txt

67 represents the max length or "width" of a line. In other words, 67 characters (including spaces).

There are other other options to choose from for wc and tee, and you can learn more about them by running man wc and man tee in Terminal.