reverse file character by character using tac

To reverse a file character-by-character using tac, use:

tac -r -s 'x\|[^x]'

This is documented in info tac:

# Reverse a file character by character.
tac -r -s 'x\|[^x]'

-r causes the separator to be treated as a regular expression. -s SEP uses SEP as the separator. x\|[^x] is a regular expression that matches every character (those that are x, and those that are not x).

$ cat testfile
abc
def
ghi
$ tac -r -s 'x\|[^x]' testfile

ihg
fed
cba%
$

tac file is not the same as cat file unless file has only one line. tac -r file is the same as tac file because the default separator is \n, which is the same when treated as a regular expression and not.

If you're not concerned about the LF character (see mosvy's comments), then it is much more efficient to use rev rather than tac -r -s 'x\|[^x]', piping it to tac if needed.

$ cat testfile
abc
def
ghi
$ rev testfile
cba
fed
ihg
$ rev testfile|tac
ihg
fed
cba

This solution is way cheaper than tac -r -s 'x\|[^x]' because regular expressions tend to slow things down quite dramatically.