Wondering what Data Denormalization means? Data Denormalization is a technique used on a previously-normalized database to increase the performance. In computing, denormalization is the process of improving the read performance of a database, at the expense of losing some write performance, by adding redundant copies of data or by grouping it. To ease up your understanding, let's go through the topic step-by-step.
Data is a set of characters generated for some purpose. They are individual units of information. Data can be in any form numbers, character, special symbols, sound, and video. The data means " information" and " knowledge". The data is composed of bits, bytes, and characters in computer language. It is processed by the CPU, which uses logical operations to bring new data from the input data. The data describe the quantity and quality also of some objects.
Denormalization? What is that?
Denormalization is a technique in which we add the duplicate data to one or more table. With the help of this, we can avoid costly joins in a relational database. Denormalization is a technique to speed up read oriented data retrieval performance in a relational database. We cannot say that when the data is not normalized, then it is in the denormalized state. What is a normalized state then?
Normalization is a technique that helps in organizing the data in the database. The main purpose of normalizing the data is to remove redundancy from the relation. Normalization divides the table into smaller tables while normalizing the table. Normalization also helps in minimizing the redundancy anomalies such as:
Updation anomaly: This anomaly occurs when the duplicated data is updated at one instance and not in the entire instance where the redundant data is present.
Deletion anomaly: This anomaly occurs when you remove the records which contain the additional important information which also gets deleted.
Insertion anomaly: This anomaly occurs when you cannot add certain attributes in the database without the presence of the other attributes.
Types of Normalization
Below mentioned are certain types of normalization which are commonly used normal forms in the database:
First normal form (1NF)
Second normal form (2NF)
Third normal form (3NF)
Boyce and Codd normal form (BCNF)
Data Denormalization and Normalization
We all now know about the normalization of the database. Denormalization does not mean that the data is not normalized. Denormalization is a technique that is performed on the normalized data. In normalization of data, we store data in separate tables to avoid redundancy due to which we have only one copy of each data in a database. In some ways, it is a good thing to happen. Why? If we update the data at one place, there will be no chance of the duplication of data. But if the no of tables is large, we have to spend much time performing joins on those tables. But with the help of Denormalization, we think that some duplicated data is okay and some efforts to perform fewer joins with the efficiency advantages. So that is why the denormalization is not the unnormalized data. Having redundant data can improve the performance in specific ways of database searches for a particular item. Know more about Data Preparation and related stuff here.
Advantages of Denormalization
Denormalization is used by the database managers to increase the performance of a database. Some of its advantages are:
Minimizing the need for joins
Reducing the number of tables
Queries to be retrieved can be simpler.
Less likely to have bugs
Precomputing derived values
Reducing the number of relations
Reducing the number of foreign keys in relation
Data modification at the computing time and rather than at the select time
Retrieving data is faster due to fewer joins.
Disadvantages of Denormalization
Although Data Denormalization can avoid some anomalies that can lead to the mismatch of the result, it may
Slow down updates, although maybe speeding up retrievals.
Make it more complex in others, although simplifying implementation.
It also can
Increase the size of relations.
Make the update and insert codes harder to write.
Involve Data redundancy which necessitates more storage.
The data can be changed now in many places, so we have to be careful while adjusting the data to avoid data anomalies. We can use triggers, transactions, or procedures to avoid such inconsistencies.
Data Denormalization vs Normalization
For all the benefits that normalizing data brings, just like anything else in information technology, there are tradeoffs and costs. A normalized relational database for even a small business could comprise hundreds of tables. For transactions, like purchases, inventory maintenance, personal data, this should not present many issues if data management is being handled through a front end application. While normalized data is optimized for entity-level transactions, denormalized data is optimized for answering business questions and driving decision making. Denormalized data is data that has been extracted from the large collection of normalized tables and has been organized and/or aggregated into fewer tables without regard to such things as redundancy. Denormalization has fewer rules about structure and not like normalization. There are schematic patterns, like Snowflake Schema, but the design is usually more specific to a particular organization’s needs. Reporting and decision support is simplified through a minimum of aggregated tables versus extracting data in real-time through multiple table joins.
In Denormalization, we add the redundant data to a normalized database so that it helps to reduce some types of problems with the database queries that combine data from different tables into a single table. It involves creating different tables or structures so that the queries when performed on those tables will not affect the information present in another table tied to it. "Do challenges arise when the business wants answers to questions like in which neighbourhoods do my best widget A customers live? The answer to such a question can drive where the business puts its advertising dollars", for example. So, now the original spreadsheet where widget A purchases are in the same record as an address isn't looking so bad! This is where denormalizing data takes centre stage.