Re: Normalized Tables & SELECT [was: Find "smallest common year"]
От | Richard Huxton |
---|---|
Тема | Re: Normalized Tables & SELECT [was: Find "smallest common year"] |
Дата | |
Msg-id | 46FCC790.7050109@archonet.com обсуждение исходный текст |
Ответ на | Re: Normalized Tables & SELECT [was: Find "smallest common year"] (Alban Hertroys <a.hertroys@magproductions.nl>) |
Ответы |
Re: Normalized Tables & SELECT [was: Find "smallest common year"]
|
Список | pgsql-general |
Alban Hertroys wrote: > If you _do_ need this table (because you want to constrain your > statistical data to only contain a specific set of years, or because you > need a quick list of available years to select from): Make the year > primary key and drop the artificial index. > > Years are perfectly fine data to constrain on, and it saves you the > joins with that table (the foreign key constraint constrains your data > sufficiently). > > You may want to constrain your years further by adding a check > constraint, fe. CHECK (year > 1900). > > The same argument goes for the country names, but I generally don't like > to have text data for primary key. Ah, but there is a standardised list of country-codes ideal for the DBA since the code is (usually) easily understandable. http://en.wikipedia.org/wiki/ISO_3166 The only problem might be if historical data uses old boundaries (e.g. Soviet Union, Yugoslavia). -- Richard Huxton Archonet Ltd
В списке pgsql-general по дате отправления: