You have a few problems. The date utility always terminates its output with a <newline>. The awk system() function sends its output to awk's standard output and returns the exit status of the command that you invoke. (It does not return the output produced by the command you invoke as a string that you can assign to a variable in awk as you tried to do in your code.)

To fix that, you can use sprintf() (in this case a couple of times) to create a date command string to give to system() that will invoke the date utility with a format string that will print the two lines of output you want each with the leading date and time field. Note that the following invokes date each time you call print_total() as in your earlier awk code:

Code:

function print_total() {
	dc = sprintf("date \"+%%a %%b %%d %%H:%%M%%S %%Z %%Y,%s-%s,%d,total%%n",
		last, last, total)
	dc = sprintf("%s%%a %%b %%d %%H:%%M%%S %%Z %%Y,%s-%s,%.1f,average\"",
		dc, last, last, total / count)
	system(dc)
}

If you just want to invoke date once at the start of your script and print that timestamp in all of the output produced by a single run of your awk script (which is closer to what you're doing now with awk and sed, but you're invoking date twice and may end up with one date and time on the total lines and a later date and time on the average lines), it would be much simpler to use:

Code:

awk -v dt="`date`" '
function print_total() {
	printf("%s,%s-%s,%d,total\n", dt, last, last, total)
	printf("%s,%s-%s,%.1f,average\n", dt, last, last, total / count)
}

PS I didn't see your last two updates before posting this. But, it should give you enough information to make changes to get what you want.