What is the difference between ' 2>&1 ' and '>'?

Question

If 2>&1 redirects the stderr to the stdout, isn't that what just '>' does? So does '2>&1' and '>' have the same meaning or are they different in some way? Explanations about redirections seem to revolve around the difference between what each of them mean but does not say if '2>&1' and '>' are the same. Can someone tell me where i am wrong ?

Answer 1

> has a meaning on its own. It is a redirection operator.

2> may be redirecting stderr, but stderr is not the only output from a task. There is also stdout which is 1 and there are also other places you could redirect to.

For example >myOutput.txt will redirect output (I believe only stdout by default) from that task to your file.

Then there is 1>stdOut.txt which explicitly redirects only stdout to the output file.

By extension you could also do 2>stdErr.txt

The reason to do 2>&1 is to redirect stderr to the same stream as stdout so that you only have one stream as an output from the command. Then that combined output can be redirected as one command.

So rather than

myCommand 1>stdOut.txt 2>stdErr.txt

You can do

myCommand 1>combinedOut.txt  2>&1 

Answer 2

So rather than

myCommand 1>stdOut.txt 2>stdErr.txt

You can do

myCommand 2>&1 1>combinedOut.txt

… it is not really correct… if you want to send stderr and stdout at the same place, the 2>&1 has to be used at the end of the command.

Here, the stderr is sent at your stdout not redirected (console), and the stdout is sent to the file with no effect on the stderr.