The discussion so far has been limited to a single disjoint data set and its sets. There can be many of these in a database. How can you be sure that you have divided the data items of your application in data sets adequately, or does it even matter? The answer is that it does matter and that you might need to make an adjustment, typically by breaking up a data set in more than one data set and sharpening the concepts that the data sets represent. This eliminates update anomalies in the process and makes the database more general and thus more amenable to future extensions. It also makes the database easier to understand.

Update anomalies can be classified in the following three types:

The three types of anomalies are discussed in the following text.