Redirecting Standard Output and Standard Error in Bash

When executing commands in Bash, programs communicate through standard streams: standard output (stdout) for normal results, and standard error (stderr) for error messages or diagnostics. Controlling where these streams go is a fundamental skill for scripting, logging, and automation. This tutorial explains how to redirect both standard output and standard error to a file, including appending to an existing file.

Understanding Standard Streams

Before diving into redirection, it’s crucial to understand the concept of file descriptors. Bash (and most Unix-like systems) uses file descriptors to represent open files. These are small integer numbers:

0: Standard input (stdin) – typically the keyboard.
1: Standard output (stdout) – typically the terminal.
2: Standard error (stderr) – typically the terminal.

Redirection operators manipulate these file descriptors to change where a command sends its output.

Redirecting Standard Output

The > operator redirects stdout to a file, overwriting the file if it exists. For example:

command > output.txt

This sends the standard output of command to output.txt. If output.txt exists, its contents are truncated (deleted).

To append to a file instead of overwriting it, use the >> operator:

command >> output.txt

This appends the standard output of command to the end of output.txt. If the file doesn’t exist, it’s created.

Redirecting Standard Error

The 2> operator redirects stderr to a file, overwriting the file if it exists.

command 2> error.txt

This sends the standard error of command to error.txt, overwriting any existing content.

Similarly, 2>> appends stderr to a file:

command 2>> error.txt

Redirecting Both Standard Output and Standard Error

Often, you want to capture both stdout and stderr in the same file. The key is to redirect stderr to the same destination as stdout.

Method 1: The Classic Approach (Portable)

This method works in all POSIX-compliant shells, including older versions of Bash. It’s the most portable and reliable approach:

command > output.txt 2>&1

Let’s break this down:

> output.txt: Redirects stdout to output.txt, creating or truncating the file.
2>&1: This is the critical part. It redirects file descriptor 2 (stderr) to the same location as file descriptor 1 (stdout). The & is essential; it tells Bash that 1 is a file descriptor, not a file named 1.

Method 2: Bash 4 and Later (Shorthand)

Bash version 4 introduced a shorthand operator for redirecting both stdout and stderr:

command &>> output.txt

This is equivalent to the classic method, but more concise. Note that this syntax is not portable to older Bash versions or other shells.

Choosing the Right Method

If portability is a concern (you need your script to run on older systems or different shells), use the classic > output.txt 2>&1 approach.
If you are confident your script will only run on Bash 4 or later, the &>> output.txt shorthand can make your code more readable.

Redirecting to Separate Files

You can also redirect stdout and stderr to different files:

command > output.txt 2> error.txt

This sends the standard output to output.txt and the standard error to error.txt. This is useful when you need to separate normal output from error messages for detailed analysis or debugging.

Example

Let’s say you have a script called my_script.sh:

#!/bin/bash
echo "This is a normal message."
echo "This is an error message." >&2

To redirect both stdout and stderr to a file called log.txt, you would run:

./my_script.sh &>> log.txt

The log.txt file would contain:

This is a normal message.
This is an error message.