If the log has a lot of repeated lines in its original state then running uniq twice, once up front to reduce what needs to be sorted, might be quicker? uniq log_file | sort| uniq | wc -l - Pad.