group_id, application_id, reading
awk is capable:
awk
awk -F "," '{print $0 >> ("FILE" $1)}' HUGE.csv