Good - let me know if you do hit any problems with it.
Quote: "What about adding Paul's linear sort method"
You mean the radix sort - I've thought about it, and I'll be looking into it. Remember though that if I implement it, it's got to work for all DBPro data types, and for multi-field arrays too:
- integer, dword, word, byte, double integer = easy
- strings = doable (variable length data)
- floats, double floats = harder but still doable
I'll have to write 4 different alternatives (the integer ones can be a single template) to do this.
All in all, a time consuming job for
potentially no benefit - apparently, the constant costs for this means it can take longer to sort than quicksort and its derivatives over small amounts of data, and for the most part, we are talking about small amounts of data. I guess I'll have to write one of the versions as a trial.
Besides, does anyone actually use the alternative sorting methods already provided anyway?