reverse a file in Unix shell

Question

I have a file parse.txt

parse.txt contains the following

remo/hello/1.0,remo/hello2/2.0,remo/hello3/3.0,whitney/hello/1.0,julie/hello/2.0,julie/hello/3.0

and I want the output.txt file as (to reverse the order from last to first)using parse.txt

julie/hello/3.0,julie/hello/2.0,whitney/hello/1.0,remo/hello3/3.0,remo/hello2/2.0,remo/hello/1.0

I have tried the following code:

tail -r parse.txt 

Answer 1

You can use the surprisingly helpful tac from GNU Coreutils.

tac -s "," parse.txt > newparse.txt

tac by default will "cat" the file to standard out, reversing the lines. By specifying the separator using the -s flag, you can simply reverse your fields as desired.

(You may need to do a post-processing step to get the commas to work out correctly, which can be another step in your pipeline.)

Answer 2

I like the tac solution; it's tight and elegant, but as Micah pointed out, tac is part of GNU Coreutils, which means that it's not available by default in FreeBSD, OSX, Solaris, etc.

This can be done in pure bash, no external tools required.

#!/usr/bin/env bash

unset comma
read foo < parse.txt
bar=(${foo//,/ })
for (( count="${#bar[@]}"; --count >= 0; )); do
  printf "%s%s" "$comma" "${bar[$count]}"
  comma=","
done

This obviously only handles one line, per your sample input. You can wrap it in something if you need to handle multiple lines of input.

The logic here is that we can convert the input into an array by replacing commas with spaces. Of course, if our input data included spaces, this would have to be adjusted. Once we have the array, we simply step backwards through it, printing each record.

Note that this does not include a terminating newline. If you want one, you can add it with:

printf '\n'

as a final line.