Greg Stark <gsstark@mit.edu> writes:
> Am I wrong? I thought the disk sort algorithm was the polyphase tape sort from
> Knuth which is always reading two tapes and writing to a third.
It is a polyphase sort, but we recycle the input "tapes" as fast as we
use them, so that the maximum disk space usage is about as much as the
data volume to sort.
regards, tom lane