Denormalization

Definition - What does Denormalization mean?

Denormalization is a strategy that database managers use to increase the performance of a database infrastructure. It involves adding redundant data to a normalized database to reduce certain types of problems with database queries that combine data from various tables into a single table. The definition of denormalization is dependent on the definition of normalization, which is defined as the process of organizing a database into tables correctly to promote a given use.

[WEBINAR] Embedded Analytics for Today's Information Workers

Techopedia explains Denormalization

In many cases, denormalization involves creating separate tables or structures so that queries on one piece of information will not affect any other information tied to it. For instance, where more global data variables such as customer names are tied together with single purchases in a purchase history, a database administrator will want to make sure that work done on an item purchased will not incorrectly affect the entire customer account. Therefore, database handlers will separate the two pieces of information, sometimes with redundant data, so that they can be worked on separately.

Where denormalization comes in is that adding redundant data allows for more sophisticated search results. Some examples that are typically given to explain this include situations where database handlers want to find prior addresses, purchase histories, or anything else about a customer or client that doesn’t address the specific present state of that account. This is where having redundant data can allow databases to give different results based on exactly what the user is asking for. Again, having this redundant data can also improve performance based on the specific ways that a database searches for a particular item. Challenges involved in denormalization include documenting the process carefully to avoid some kinds of anomalies that can occur as a result of data mismatch.