SQL Antipatterns, Volume 1

Legitimate Uses of the Antipattern

FLOAT is a good data type when you need real number values with a greater range than INTEGER or NUMERIC data types support. Scientific applications are often the best use of a FLOAT. For example, if you store daily temperature data in a database, the fractional values will vary, and will seldom be a whole degree. It’s okay to store a temperature value with a billionth of a degree, because the common use of those data is to make aggregate calculations, like MIN(), MAX(), or AVG(). You might also search for rows where the temperature is in a range of values, and a query using inequality or BETWEEN will work fine with inexact numerics. But you’re not likely to search for the row where the temperature is exactly ...