1390

How do I save the output of a command to a file?

Is there a way without using any software? I would like to know how.

0

11 Answers 11

1715

Yes it is possible, just redirect the output (AKA stdout) to a file:

SomeCommand > SomeFile.txt  

Or if you want to append data:

SomeCommand >> SomeFile.txt

If you want stderr as well use this:

SomeCommand &> SomeFile.txt  

or this to append:

SomeCommand &>> SomeFile.txt  

if you want to have both stderr and output displayed on the console and in a file use this:

SomeCommand 2>&1 | tee SomeFile.txt

(If you want the output only, drop the 2 above)

15
  • 40
    Note that someCommand 2> someFile.txt and someCommand 2>> someFile.txt also redirects stterr to someFile.txt
    – Slothworks
    Aug 29, 2015 at 13:32
  • I'm trying to do this with gcc command but it doesn't work. It works with other commands, but not this one. It simply creates the output file with nothing inside it.
    – KeyC0de
    Sep 29, 2016 at 13:40
  • @Nik-Lz Often this is because the command is sending all its output on stderr. If gcc is generating error messages, this seems likely. See Slothworks comment for how to capture stderr instead of stdout. Sep 14, 2017 at 13:29
  • 1
    NB: to get the output of the make command into a file it requires this syntax instead: make > someFile.txt 2>&1 (source: linuxquestions.org/questions/linux-newbie-8/…) Jan 15, 2018 at 0:49
  • 1
    I have a problem that it stops writing when the file reaches about 8MB. Is this a known limit?
    – relG
    Oct 27, 2018 at 8:02
1580

To write the output of a command to a file, there are basically 10 commonly used ways.

Overview:

Please note that the n.e. in the syntax column means "not existing".
There is a way, but it's too complicated to fit into the column. You can find a helpful link in the List section about it.

          || visible in terminal ||   visible in file   || existing
  Syntax  ||  StdOut  |  StdErr  ||  StdOut  |  StdErr  ||   file   
==========++==========+==========++==========+==========++===========
    >     ||    no    |   yes    ||   yes    |    no    || overwrite
    >>    ||    no    |   yes    ||   yes    |    no    ||  append
          ||          |          ||          |          ||
   2>     ||   yes    |    no    ||    no    |   yes    || overwrite
   2>>    ||   yes    |    no    ||    no    |   yes    ||  append
          ||          |          ||          |          ||
   &>     ||    no    |    no    ||   yes    |   yes    || overwrite
   &>>    ||    no    |    no    ||   yes    |   yes    ||  append
          ||          |          ||          |          ||
 | tee    ||   yes    |   yes    ||   yes    |    no    || overwrite
 | tee -a ||   yes    |   yes    ||   yes    |    no    ||  append
          ||          |          ||          |          ||
 n.e. (*) ||   yes    |   yes    ||    no    |   yes    || overwrite
 n.e. (*) ||   yes    |   yes    ||    no    |   yes    ||  append
          ||          |          ||          |          ||
|& tee    ||   yes    |   yes    ||   yes    |   yes    || overwrite
|& tee -a ||   yes    |   yes    ||   yes    |   yes    ||  append

List:

  • command > output.txt

    The standard output stream will be redirected to the file only, it will not be visible in the terminal. If the file already exists, it gets overwritten.

  • command >> output.txt

    The standard output stream will be redirected to the file only, it will not be visible in the terminal. If the file already exists, the new data will get appended to the end of the file.

  • command 2> output.txt

    The standard error stream will be redirected to the file only, it will not be visible in the terminal. If the file already exists, it gets overwritten.

  • command 2>> output.txt

    The standard error stream will be redirected to the file only, it will not be visible in the terminal. If the file already exists, the new data will get appended to the end of the file.

  • command &> output.txt

    Both the standard output and standard error stream will be redirected to the file only, nothing will be visible in the terminal. If the file already exists, it gets overwritten.

  • command &>> output.txt

    Both the standard output and standard error stream will be redirected to the file only, nothing will be visible in the terminal. If the file already exists, the new data will get appended to the end of the file..

  • command | tee output.txt

    The standard output stream will be copied to the file, it will still be visible in the terminal. If the file already exists, it gets overwritten.

  • command | tee -a output.txt

    The standard output stream will be copied to the file, it will still be visible in the terminal. If the file already exists, the new data will get appended to the end of the file.

  • (*)

    Bash has no shorthand syntax that allows piping only StdErr to a second command, which would be needed here in combination with tee again to complete the table. If you really need something like that, please look at "How to pipe stderr, and not stdout?" on Stack Overflow for some ways how this can be done e.g. by swapping streams or using process substitution.

  • command |& tee output.txt

    Both the standard output and standard error streams will be copied to the file while still being visible in the terminal. If the file already exists, it gets overwritten.

  • command |& tee -a output.txt

    Both the standard output and standard error streams will be copied to the file while still being visible in the terminal. If the file already exists, the new data will get appended to the end of the file.

28
  • 107
    Thanks for the table, it's excellent! This should be top answer
    – DevShark
    Aug 15, 2016 at 16:24
  • 3
    @karthick87 This is not really related to the question about redirecting output to a file, because it just redirects one stream to another. 2>&1 redirects STDERR to STDOUT, 1>&2 redirects STDOUT to STDERR and 3>&1 would redirect stream 3 to STDERR.
    – Byte Commander
    Sep 19, 2016 at 16:42
  • 25
    Just a note that '|&' wasn't working for me on macOS. This is due to it having an older version of bash (I think). The less elegant '2>&1 |' works fine though May 9, 2017 at 10:01
  • 2
    @ByteCommander I get the error: sh: 1: Syntax error: "&" unexpected when I use |& tee from a Python script in a c9.io server. It seems a different shell is being used. echo $SHELL shows /bin/bash and $SHELL --version shows version 4.3.11(1)-release. I tried #!/bin/bash in my python script but I still get sh: 1: Syntax error. I got what I needed so I'm giving up on sorting the weirdness between sh and bash on my server. Thanks.
    – samkhan13
    Jan 28, 2018 at 3:09
  • 1
    @samkhan13 looks like you are running sh and not bash (or maybe bash in sh mode...). You can check what exactly your current shell process is using ps -p $$ -o cmd=, because echo $SHELL is unreliable and will show you your login shell, ignoring whether you might have started a different subshell.
    – Byte Commander
    Jan 28, 2018 at 12:35
138

You can also use tee to send the output to a file:

command | tee ~/outputfile.txt

A slight modification will catch stderr as well:

command 2>&1 | tee ~/outputfile.txt

or slightly shorter and less complicated:

command |& tee ~/outputfile.txt

tee is useful if you want to be able to capture command output while also viewing it live.

10
  • It says that the & is unexpected, and doesn't write the log at the same time as the command runs. I am using this in a bash file however, does that make any difference?
    – tim687
    Apr 6, 2016 at 13:51
  • @tim687 I have removed that edit. Sorry about that...wasn't a part of my original answer.
    – Aaron
    Apr 6, 2016 at 14:11
  • 1
    how do I interpret meaning of 2>&1?
    – Mahesha999
    Jul 15, 2016 at 7:47
  • 3
    @Mahesha999 2 is the file descriptor for STDERR, and 1 is for STDOUT. So that 2>&1 sends STDERR to STDOUT. This SO question explains it pretty well: stackoverflow.com/questions/818255/…
    – Aaron
    Jul 15, 2016 at 7:53
  • 1
    Instead of 2>&1 |, one can also simply use |&.
    – Byte Commander
    Oct 7, 2016 at 17:47
24

The script command

There are two different questions here. The first is in the title:

How do I save terminal output to a file?

The second question is in the body:

How do I save the output of a command to a file?

All the answers posted here address the second question but none address the first question which has a great answer in Unix & Linux:

This answer uses a little known command called script which saves all your shell's output to a text file until you type exit. The command output still appears on your screen but also appears in the text file.

The process is simple. Use:

$ script ~/outputfile.txt
Script started, file is /home/rick/outputfile.txt
$ command1
$ command2
$ command3
$ exit
exit
Script done, file is /home/rick/outputfile.txt

Then look at your recorded output of commands 1, 2 & 3 with:

cat ~/outputfile.txt

This is similar to earlier answer of:

command |& tee ~/outputfile.txt
  • But you don't have to use |& tee ~/outputfile.txt after each commnd.
  • The script command has added benefit (or disadvantage) of reloading ~/.bashrc when it starts.
  • The script command shows the command prompt ($PS1) followed by the command(s) you entered.
  • The script command records all the details in full color.

Send output to clipboard

Many times we want the output to go to the clipboard so we can paste it later. From this answer you can use:

cat ~/.bashrc | xclip -selection clipboard

Now you can use Ctrl+V in almost any application to paste the terminal output into your document. To paste the terminal output in the clipboard back into your terminal use Ctrl+Shift+V instead.

5
  • 2
    +1 :-) I use script in the following context: Wake me up when a slow command line process wants my attention?
    – sudodus
    Dec 31, 2019 at 0:09
  • 2
    @sudodus +1 on your linked answer there. I would never have dreamed of using script for such an application :) Dec 31, 2019 at 0:16
  • This looks promising but didn't work for me with Cygwin. Script started on 2021-08-20 11:14:49+10:00 [TERM="xterm" TTY="/dev/cons0" COLUMNS="80" LINES="24"] (script is empty) Script done on 2021-08-20 11:14:49+10:00 [COMMAND_EXIT_CODE="0"]
    – Greg
    Aug 20, 2021 at 9:04
  • 2
    script is the solution you need when you have enter interactive responses y/n etc
    – zzapper
    Oct 12, 2021 at 17:11
  • @Greg This is a Ubuntu site. Your problem is with Cygwin which is used to run Linux on Windows. To run it in Ubuntu you would have to use wine. So you would be using Linux to Cygwin to Windows to Wine to Linux for your problem to be supported here on Ask Ubuntu. Otherwise it would be a problem best posted on Super User. Dec 13, 2021 at 1:10
23

You can redirect the command output to a file:

your_command >/path/to/file

To append the command output to a file instead of overwriting it, use:

your_command >>/path/to/file
5
  • Thanks a lot ! is there any Limits ? like the max size of the file ?
    – led-Zepp
    Feb 14, 2014 at 19:55
  • 3
    The max file size is just limited through the file system
    – chaos
    Feb 14, 2014 at 19:59
  • This answer will not save stderr. Use &> , see stackoverflow.com/questions/637827/… and tldp.org/LDP/abs/html/io-redirection.html
    – Panther
    Feb 14, 2014 at 20:16
  • 3
    The OP never asked to save stderr
    – chaos
    Feb 14, 2014 at 20:18
  • It says "No such file or directory". Is it possible to also create the directories automatically?
    – Qwerty
    May 22, 2019 at 9:27
22

An enhancement to consider -

Various scripts will inject color codes into the output which you may not want cluttering up your log file.

To fix this, you can use the program sed to strip out those codes. Example:

command 2>&1 | sed -r 's/'$(echo -e "\033")'\[[0-9]{1,2}(;([0-9]{1,2})?)?[mK]//g' | tee ~/outputfile.txt
4
  • 2
    How to save the output in a way that colours are conserved ? I would like to import the result of a command in libreoffice and keep the colours.
    – madrang
    May 12, 2015 at 6:36
  • 1
    @madrang: I only read your comment now but you may find this answer useful. Sep 21, 2015 at 10:41
  • Oh, almost exactly what I am looking for. How to print also on screen the output?
    – Sigur
    Dec 23, 2016 at 22:10
  • 1
    Note that many commands that produce colorized output, such as ls and grep, support --color=auto, which outputs color codes only if standard output is a terminal. Sep 3, 2017 at 14:37
15
+50

some_command | tee command.log and some_command > command.log have the issue that they do not save the command output to the command.log file in real-time.

To avoid that issue and save the command output in real-time, you may append unbuffer, which comes with the expect package.


Example:

sudo apt-get install expect
unbuffer some_command | tee command.log
unbuffer some_command > command.log

Assuming log.py contains:

import time
print('testing')
time.sleep(100) # sleeping for 100 seconds

you can run unbuffer python log.py | tee command.log or unbuffer python log.py > command.log

More information: How can I save a command output to a file in real-time?

3
  • 1
    They do save the output as they receive it, the problem is that python turns on buffering when the output is not to a TTY. Other options for disabling this in Python: stackoverflow.com/q/107705/2072269
    – muru
    Jul 5, 2018 at 2:22
  • 1
    Thanks! Spent so long looking for this, and this is exactly what I needed. Works with the stdout from the Google Assistant scripts.
    – anonymous2
    Dec 26, 2019 at 2:23
  • You don't need to install expect, stdbuf does the same thing and is generally already there in your distribution. You can "line buffer", which still has the advantage of some buffering and saves time, or completely "unbuffer".
    – Zakhar
    Aug 23, 2020 at 18:06
11

For cron jobs etc you want to avoid the Bash extensions. The equivalent POSIX sh redirection operators are

Bash          POSIX
------------  --------------
foo &> bar    foo >bar 2>&1
foo &>> bar   foo >>bar 2>&1
foo |& bar    foo 2>&1 | bar

You'll notice that the POSIX facility is in some sense simpler and more straightforward. The &> syntax was borrowed from csh which should already convince you that it's a bad idea.

3

If you want to output to the file while the command is being run:

script -c ./path/to/executable.bash -f log.txt
3

Use terminal emulator features

An option not mentioned yet, that can save colours / colors too, is to use a console program — such as Konsole (KDE/Plasma's default terminal emulator) — to save the output.

Konsole

Konsole has: 
File > Save output as... 

the shortcut is Ctrl+Shift+S; it allows the output to be saved as a text file, or as HTML including colors! I'm not sure exactly how much it will save but in my test it only included ~1000, the entire terminal scrollback, (you can increase the buffer by creating a new profile, Profile > New..., and then change the Scrolling settings to capture more; Konsole version 4:21.08.1).

gnome-terminal

gnome-terminal has "copy output as HTML" too, which allows pasting the HTML into a document; it preserves colour but only copies the content of the output currently shown on screen AFAICT.

generically

You can, of course, do a straight drag-select (hold left mouse button whilst dragging) and then copy (ctrl+c, Edit > Copy, or right-mouse-click and choose copy).

others?

Feel free to modify this answer to include other popular terminal apps. My favourite, Yakuake, does not appear to have this feature nor did most of the popular terminals I reviewed, including terminal.app and Hyper.

0

Add this to your .bashrc (based on the script command):

### Logging function:
start_terminal_logging()
{   
    if [ -z "$SCRIPT_LOG_FILE_ACTIVE" ]; then    
        # Log each session to a separate file:
        logdir=~/.terminal_logs
        logfile=$logdir/session_log_$(date +%Y%m%d_%H%M%S)_$$.txt        
        # If no folder exist make one:
        if [ ! -d $logdir ]; then
            mkdir -p $logdir
        fi
        export SCRIPT_LOG_FILE_ACTIVE=$logfile
        # Start logging, a=append, q=quiet:
        script -aq $logfile
        exit
    fi
}
### Start logging:
start_terminal_logging

This will give you a separate file for each session:

session_log_20240107_152455_1955.txt
session_log_20240124_112455_2971.txt
session_log_20240125_092455_6925.txt
session_log_20240125_092750_7048.txt
session_log_20240125_092808_7079.txt
.
.
.

Filename contains a timestamp, so you can easier find what you are looking for.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .