As far as I can tell, the SQL Server gurus out here in the interwebs seem to be in general agreement that there are performance gains to be had by using a two-part table naming convention in your SQL. For example:
SELECT shirt.sleeve_length
FROM dbo.clothes shirt
Instead of
SELECT shirt.sleeve_length
FROM clothes shirt
(Also noting that when using linked servers or other dbs on the same server, you can get back to a two-part convention using synonyms)
Given the calibre of the commentators who espouse this view, there's a pretty good chance they're right :-)
My "shop" does not subscribe to this view though, because when they write their queries in SSMS and hit F5 there is no discernible difference in run time. "...so why should we waste our time adding the schema to everything?" I don't think their experimental design is particularly empirical, but I don't have an empirical design myself, that will provide evidence to the contrary.
Is there a way I can demonstrate empirically that a two-part table naming convention is better?
↧