A SELECT query returns all the rows for which the WHERE clause returns True. However, many developer - especially those accustomed to other programming languages, such as VB - get confused on this point, and assume that the query returns the rows for which the WHERE clause returns any non-False value, and this is a major source for bugs.
To explain the reason, consider that the SQL language uses a three-value logic: True, False, and Null. Any Null value in an expression makes the entire expression Null (with some exception, noted below). For example, your common sense would suggest that the following SELECT returns all the rows in the table:
SELECT * FROM Orders WHERE (total < 1000 Or total >= 1000)
What happens, however, is that the SELECT won't include records for which the Total field is Null, because this value makes the entire WHERE expression evaluate to Null. Here's another example:
SELECT * FROM Orders WHERE total = 0
The above query returns all orders for which Total is zero, but not those for which Total is Null. If you want to include the latter ones, you must explicitly look for Null values:
SELECT * FROM Orders WHERE total = 0 OR total IS NULL
ISNULL is T-SQL a function that is often useful to get rid of Null values. Simply stated, it always returns its first argument, except when it is Null (in which case it returns the second argument). See how you can rewrite the above query to avoid Null values:
-- convert Null values to zero before comparing them
SELECT * FROM Orders WHERE ISNULL(total, 0) = 0
Moreover, T-SQL extends the ANSI 92 standard and supports Null also in IN clauses, so you can rewrite the above query as follows:
SELECT * FROM Orders WHERE total IN (0, NULL)
