"do not do this because it is bad" - which it is not necessarily. In addition, that phrase might also be interpreted as In the "old" days it was far more common since this was a technique to accommodate the same concept for which we use nullable columns now. Kalman gave you a link to BOL, but I would argue that the use of the phrase "is notĬommon" is rather unfortunate and even a bit misleading. If we assume that you have good reasons for partitioning the table (and you haven't really provided any, hence the assumption) then the logical choice is the first. Yes, this is also done and is typically used to support a 1:M relationship between the primary table and the dependent table. The only difference in this 2nd question is that you propose to createĪn artificial primary key for each table and add another column to be used as the foreign key to the primary table. Your first question every table will have a primary key but all of the dependent tables will also have a foreign key (using the same column as the primary key) to the primary table. "shall I give each table it's own primary key."? Well, we have a problem with this statement due to your misuse of the terminology (though we can see through that to understand your question). "? It can be and there is nothing about this that violates any normal form. IMO some your responses have not been particularly helpful.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |