Re: Question on Structuring Product Attributes

Now you are obligated to actually post better techniques and explain why they are better..

I am not bothered by increasing the number of indexes. You have made the assumption that references are done by more indexes. Sybase (nee WATCOM) SQL implements the constraints like  FOREIGN KEY (vin, vehicle_type)
  REFERENCES Vehicles(vin, vehicle_type)

by building pointer chains under the covers, so each compound key actually appears once in the schema. Teradata uses a hashing algorithm. RDBMS companies have been working on PK-FK joins for a few decades, so they are pretty good at it now.

What do you have against the Nested Sets? The only laugh I had with them was teaching them to XML programmers. Traditional procedural programmers still think in terms of fake pointer chains, recursive traversals and other stiff they know from assembly language. But the XML kids look at it and say "Oh, (lft, rgt) is the tag pairs!" then wrote really great code by mimicking XQuery. Received on Sat Oct 15 2011 - 04:29:27 CEST