What is this syntax at the end of a while loop?

Question

In the following BASH script code:

while read -r line
do
    if [some_condition]; then
        do_thing1
        do_thing2
        continue
    fi
done < $SOME_VAR

What is the meaning of the last line? Specifically, what is being done with the value of $SOME_VAR?

Answer 1

The < is the input redirection operator, $SOME_VAR probably contains the name of a file. Together, they'd cause the loop to read lines from the file.

Answer 2

< is the shell indirection operator. It takes the provided file, and uses it as standard input for the preceding command. That command would be roughly analagous to:

cat $SOME_VAR | while read -r line; do
  if stuff; then
     consequence
  else
     otherstuff
  fi
done

If SOME_VAR were not assigned to a file in the current working directory (or the full path to a file) the shell would throw a No such file or directory error.

Answer 3

The other two answer's already covered most aspects of it. I would like to add that, to understand this syntax you should interpret the while ... done as a single but complex statement, and $SOME_VAR is fed to this statement as stdin. It will be consumed whenever there is a need to read from stdin. In this case the read statement. There is actually no difference with the standard cmd <in >out syntax.

Answer 4

In a very short terms, < redirects file specified to the right of the < symbol into stdin stream of whatever command or group of commands you're executing.

The while read -r line; do...done < input.txt is standard way to read a file line by line. While, of course, $SOME_VAR doesn't have to be text file - can be named pipe among other things - this is the usual way and is used to avoid problems with word splitting in shell.

Often new to shell scripting think they can do something like for line in $(cat file.txt) but if each line has multiple words on it separated by spaces, the for loop will be processing each word instead of each line. Something like this:

$ for i in $(cat input.txt); do echo "$i" ; done
one
two
three
four

That's not good. By contrast, while loop lets you process whole line without issues.

$ while read -r line; do echo "$line" ;done < input.txt
one two
three four

You might be wondering why not just use cat | while read -r line..., but that's extra process and useless use of cat.

---

Redirection into stdin stream of a loop can be done with other things as well. If you are using bash that supports <<<, you could send a string literal of text to your while loop.

while read -r line;do
    # do something with line
done <<< "Hello World"

A redundant example ? Yes, but proves the point that such syntax is more about redirecting input to while loop.

See also:

• http://mywiki.wooledge.org/DontReadLinesWithFor
• http://mywiki.wooledge.org/BashFAQ/001