test it. There's no way anyone can reply to your question without a db backend, a data model, the cardinality of your tables and the queries you'll most likely do. In any case, a unique-integer-serial is most of the times the "less expensive" way to hook up tables by references (indexes and "pointers" to the row using integers fit best in memory, are easier to update and they tend to show less fragmentation)
On Monday, April 22, 2013 8:08:31 AM UTC+2, Alex Glaros wrote: > > The data modeler in me prefers to use the natural key in a table, e.g., > the abbreviation for a state like CA for California, NY for New York, > instead of the web2py default surrogate integer primary key. > > > > Is there a speed advantage in a very large site with lots of users and > joins, to using the default web2py key for all tables? E.g., using the > web2py default key as the foreign key in all tables and also using it for > all “requires” and “represents” functionality? How much of a difference? > > > > Thanks, > > > Alex Glaros > -- --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
