If you are reading a file line-by-line as suggested by MadeInGermany in post #2, please never feed those lines through a three element pipeline. Doing so is GROSSLY inefficient!

If you are processing multiple files (which would be a logical reason for adding the name of the file each record came from in the output you produce), you can easily modify MadeInGermany's shell suggestion from post #2 with something like:

Code:

for file in file*.txt
do	while IFS="" read row
	do	case $row in
		(*" setting application "*)
			case $row in
			(*"@msad"*)
				;;
			(*)	printf "%s %s\n" "$row" "$file";;
			esac;;
		esac
	done < "$file"
done > tempfile.txt

(which just uses shell built-ins and doesn't need to invoke any external utilities) or extend my awk suggestion from post #3 to something like:

Code:

awk '/ setting application / && ! /@msad/ { print $0, FILENAME }' file*.txt > tempfile.txt

(which just invokes awk once). Note the comma between $0 and FILENAME in the print statement; without it, there won't be any separation between the input lines and the name of the file from which it was extracted.

The code that you showed us in post #4 invokes two copies of grep and one copy of awk for every line read from each of the files you process.