On Wed, Jan 04, 2006 at 09:12:42PM +0200, Krasimir Angelov wrote:
The point isn't in the amount of coding but in the performance. It isn't required to build intermediate data structures.
Well, you've got intermediate data structures in HSQL. In particular, each column access must traverse an association list based on the name of the column. I actually expect HDBC to be faster in this regard. It may be fun to run some benchmarks next week. I think you are saying that HSQL could be faster because it calls down into the database for each specific column reference at that time. Perhaps so, but with lists as they are in Haskell -- very inexpensive to create and traverse -- I doubt that there is a significant penalty from doing it the HDBC way, even if compared to lookups by column index in HSQL. -- John