On Fri, Jul 13, 2007 at 03:29:16PM +0100, Gregory Stark wrote:
> The key to the algorithm is that it uses a trie to bin rows with common
> leading prefixes together. This avoids performing redundant comparisons
> between those columns later.
Sounds like a variation on the idea suggested before, which is to allow
each datatype to provide an xfrm function that returns a signed
integer, which would allow you to compare values without invoking the
actual datatype comparison function in most cases. That approach would
work on any datatype, not just strings.
Whether it's more efficient than the current method is another question
entirely.
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> From each according to his ability. To each according to his ability to litigate.