I currently have an awk that removes all new lines and works fine, but I'd like to do it in a sed file since everything else is happeing in the sed file. in the sed file I have $!N;s/\n// but it only removes every other new line.

Yes, your solution pulls line2 into the buffer with line1. With two lines in the buffer, sed can "see" and operate on the embedded newline that separates the two lines, but not the newline that inherently ends the buffer. After it finishes processing your command (deletion of the newline), it writes out the now joined line1line2 which is terminated by a newline, then reads line3 into the empty buffer and executes your set of commands again. For sed to delete ALL newlines, I create a loop that keeps pulling all lines into the buffer. Here are two slightly different versions of that: sed -e :a -e '$!N;s/\n//;ta' myfile sed -e :a -e '$b;N;s/\n//;ba' myfile