Don Marvick <donmarvick(at)gmail(dot)com> writes:
> Since nobody replied, I would give it a try. I am going to implement the
> merge pattern described in Knuth Page 365 (5.4.9), essentially it is as
> follow:
> - create initial runs using replacement selection (basically this is as in
> the current implementation)
> - add enough dummy runs of size 0 so that the number of sorted run minus one
> can be divided by k-1 (k is merge fan in)
> - repetitively merge k smallest runs until only 1 run left
How is this better than, or even different from, what is there now?
regards, tom lane