Author: Dann Corbit
Date: 13:42:47 03/03/06
Go up one level in this thread
On March 03, 2006 at 16:19:00, Mark Boylan wrote: >On March 03, 2006 at 15:24:26, Steve Maughan wrote: > >> >>1) Use a hash code based system i.e. 'good' positions are stored as their hash >>table > > >Would this mean that every program would need to use the same hash values? Suggestion: Have a standard, simple hash that eats an EPD record and spits out an 8 byte hash. This (then) will be the same for every program. The input to find a position in the book will be the EPD string, which is the same for every case. In case a hash is not unique (a real possibility with hundreds of millions of records) add an additional byte to the end of the hash field. When inserting into the position table, check to see if the hash already exists. If the hash does exist, check to see if the EPD record is identical. If it is identical, then we update (SQL MERGE operation). If it is different, we add 1 to the extra byte and repeat. If we run out of hash records and we did not find our record, then we insert it. This could be handled by a SQL trigger (or a stored procedure).
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.