De-normalization in SQL

Facts are raw data that are stored in databases. There are two ways of storing data. One is one fact at one place and another one is one fact is scattered at different places.

Storing at one place is normalization and to get all the related facts we need to join all the tables. But, while doing modifications this type of database systems will be very useful. But, if we need to join many tables the performance of the query while retrieving data gets degraded.

Storing same fact at multiple places is denormalization and this type of architecture supports faster querying of the data. But, at the time of modification the performance will be very poor.

Most of the times denormalization principle is applicable where you need to query data more than with minor modifications. Like data warehousing and date mining where reporting of the data should be faster. Most of the business users and CEOs of the companies would like to analyze data based on the facts like salaries per year or number of employees joined per month in each year or any figures that helps in making decisions by the management.

2 Replies

Know the answer? Post it — somebody with the same question will find it here.

Satyapriya Nayak14y ago
http://en.wikipedia.org/wiki/Denormalization
Akkiraju Ivaturi14y ago
Facts are raw data that are stored in databases. There are two ways of storing data. One is one fact at one place and another one is one fact is scattered at different places.

Storing at one place is normalization and to get all the related facts we need to join all the tables. But, while doing modifications this type of database systems will be very useful. But, if we need to join many tables the performance of the query while retrieving data gets degraded.

Storing same fact at multiple places is denormalization and this type of architecture supports faster querying of the data. But, at the time of modification the performance will be very poor.

Most of the times denormalization principle is applicable where you need to query data more than with minor modifications. Like data warehousing and date mining where reporting of the data should be faster. Most of the business users and CEOs of the companies would like to analyze data based on the facts like salaries per year or number of employees joined per month in each year or any figures that helps in making decisions by the management.