Quicksort should be the first sort you try on your rando dataset. The native search implementation is nearly always competitive with the best algorithm for your particular lunch. And it can be tweaked to specialize on datasets.