In the above example 'muspi merol' (the output of the first rev command) is sent to stderr and 'lorem ipsum' (the output of the second rev command) is sent to stdout. rev reverse lines of a file or files. This use of tee allows testing if a program correctly handles its input without using files that hold the data.

This will get the job done in the most efficient way -

spawning only one `rm` process.

"On-the-fly" find data is displayed through `tee` and

you should have plenty of time to ctrl-c if needed before it's too late.

You may need to re-run this after major Software Updates.

To leave more languages in, add more ``-and \! -iname "lang*"'' statements:

**Edit: note the 2nd sudo near the end of the pipeline - this is necessary.

only for sudo-style systems.

Use this construct instead of I/O re-directors ``>'' or ``>>'' because

sudo only elevates the commands and *not* the re-directors.

***warning: remember that the `tee` command will clobber

file contents unless it is given the ``-a'' argument

Also, for extra security, the "left" command is still run unprivileged.

[re]verify those burned CD's early and often - better safe than sorry -

at a bare minimum you need the good old `dd` and `md5sum` commands,

but why not throw in a super "user-friendly" progress gauge with the `pv` command -

adjust the ``-s'' "size" argument to your needs - 700 MB in this case,

and capture that checksum in a "test.md5" file with `tee` - just in-case for near-future reference.

*uber-bonus* ability - positively identify those unlabeled mystery discs -

for extra credit, what disc was used for this sample output?

The large context number (-C 1000) is a bit of a hack, but in most of my use cases, it makes sure I'll see the whole log output.

Useful for cron jobs -- all output will be logged but only errors will cause email to be sent. NB the order of "2>&1" and ">> logfile" is important, it doesn't work if you reverse them (everything goes to the logfile, nothing left for tee).

Forwards localhost:1234 to machine:port, running all data through your chain of piped commands. The above command logs inbound and outbound traffic to two files.

Tip: replace tee with sed to manipulate the data in real time (use "sed -e 's/400 Bad Request/200 OK/'" to tweak a web server's responses ;-) Limitless possibilities.

This is the solution to the common mistake made by sudo newbies, since

does NOT add to the file as root.

Alternatively,

should be replaced by

And you can add a >/dev/null in the end if you're not interested in the tee stdout :

Using process substitution, we can 'trick' tee into sending a command's STDOUT to an arbitrary number of commands. The last command (command4) in this example will get its input from the pipe.

Find all files that contain string XXX in them, change the string from XXX to YYY, make a backup copy of the file and save a list of files changed in /tmp/fileschanged.

The tee (as in "T" junction) command is very useful for redirecting output to two places.