Im trying to implement a high-level quicksort algorithm using a specialized database object using tFilestream. it seems to work ok until i get to about 600,000 records, then i get stack overflows. mostly this is all upramping from taking the whole database and creating an indexed field to sort in a textfile, then running a batchfile "sort index.txt > index_sort.txt".
basically, i'd like to see it run as fast or faster than good ol' DOS sort. any ideas?