Question

I want to put a lot of filtering in a Bash script, to get my filtering commands a little shorter:

#before:
$ cat unreadable.log | grep "bla" |[...]| sed "s/blubb//" >>readable.log
#after:
$ cat unreadable.log | the_script_I_m_writing.sh >>readable.log

But how to handle the IO in the bash-script correctly, that the after line can be executed as shown?
As you can see inside the bash-script there will be a lot of simple commands that are piped together. All this piping should just begin with stdin and send the final result out to stdout.

Answer 1

cat reads from standard input if no file argument is given.

#!/usr/bin/env bash
cat | grep "bla" | sed 's/blubb//'

---

Alternatively, just write a script that accepts the source file name as argument.

#!/usr/bin/env bash
cat "$1" | grep "bla" | sed 's/blubb//'

Use: ./script unreadable.log.

---

In this simple case, you can also skip the cat, and just start with grep.

#!/usr/bin/env bash
grep "bla" | sed 's/blubb//'

Use:

# unnecessary cat if it's just a file
cat file | ./script.sh

# get stdin from file
./script.sh < file