On Thu, 22 Jan 2015 13:40:56 -0800, Andrei Alexandrescu wrote: > There's this classic patter on Unix: |sort|uniq, i.e. sort some data and > only display the unique elements. > > What would be a better integrated version - one that does sorting and > uniq in one shot? I suspect the combination could be quite a bit better > than doing the two in sequence. > > A few google searches didn't yield much. Ideas? > > > Thanks, > > Andrei
Efficiency improvement-wise, perhaps a generalization of a counting sort (http://en.wikipedia.org/wiki/Counting_sort), see "Variant algorithms."
