Flags inside sed command file

I have a sed file that replaces all occurrences of a string in a file with other string. I want to do it inline but without using -i from terminal What changes are to be made to the .sed file

#!/bin/sed
s/include/\#include/ 

2 Answers

Just use awk:

{ sub(/include/,"#include"); rec = rec $0 RS }
END{ printf "%s", rec > FILENAME }

or if you want to operate strictly on strings:

BEGIN{ old="include"; new="#include" }
s = index($0,old) {$0 = substr($0,1,s-1) new substr($0,s+length(old))
{ rec = rec $0 RS }
END{ printf "%s", rec > FILENAME }

which can be simplified to:

s = index($0,"include") {$0 = substr($0,1,s-1) "#" substr($0,s)
{ rec = rec $0 RS }
END{ printf "%s", rec > FILENAME }

in this particular case of just prepending # to a string.

I don't think it will work, because the -i and -f options can usually both have arguments, but you could be lucky.

The shebang line can contain one option group (aka cluster). You would need for it to contain -f (this is missing from your example) so the cluster could look like

#!/bin/sed -if

provided that your dialect doesn't require an argument to the -i option, and permits clustering like this in the first place (-if pro -i -f).

The obvious workaround is to change the script to a shell script; the -f option is no longer required because the script is not in a file.

#!/bin/sh
exec sed -i 's/include/\#include/' "$@"