3 Missing Values

Of the thirty-six alternatives, running away is best.—Chinese Proverb

Missing values and the proper handling of them is a very delicate subject within the database community. The debate centers on how (or whether) missing values should be stored in relational databases and how many and what types of tokens should be used to represent them in SQL.

There are at least three different schools of thought regarding how to handle missing values. The inventor of the relational database, Dr. E. F. Codd, advocates two separate missing value tokens: one for values that should be there but aren’t (e.g., the gender of a person) and one for values that shouldn’t be there at all because they are inapplicable (e.g., the gender of a corporation). ...

Get The Guru's Guide to Transact-SQL now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.