Three paradigms of database

This article is about Shangsi valley MySQL Notes written

1. Why database design is needed

When designing data tables , There are many problems to consider :

• What data do users need , What data do we want to save in the data table
• How to ensure the correctness of the data in the data table
• How to reduce the redundancy of data table
• How can developers use databases more conveniently

If the database design is unreasonable , It may cause the following problems :

• Easy to design , Repetition of information , Waste of storage space
• Data update , Insert , Deleted exception
• Does not represent information correctly
• Loss of valid information
• Poor program performance

We can see that a well-designed database is very important , It has the following advantages :

• Save data storage space
• It can guarantee the integrity of data
• Convenient for the development of database application system

Design database , We have to pay attention to the design of data table , In order to create small redundancy , A well structured database , Designing a database must follow certain rules .

2. normal form (Normal Formal）

2.1 Paradigm overview

In relational database , About the basic principles of data table design , Rules are called paradigms , Normal form is the rule and guiding method that we need to follow in the process of designing database structure .

however , Sometimes in order to improve some query performance , We also need to break the paradigm rules , It's anti normalization .

2.2 The concept of keys and related attributes

The definition of the normal form uses primary keys and candidate keys , Let's first look at the related concepts , A key in a database is made up of one or more attributes , Let's take a look at the definitions of several keys and attributes commonly used in data tables .

2.3 First normal form (1NF)

• The first normal form mainly guarantees that the value of each field in the data table must have Atomicity , That is, the value of each field in the data table is It can't be broken up any more The smallest data unit
• 
• 
• The atomicity of attributes is subjective , We should design according to the needs of the actual project , Like the address , If the project does not say to be subdivided into provinces , City , county , If the town is so specific , Generally, we can not split it .

2.4 Second normal form (2NF）

• The second paradigm requires that on the basis of meeting the first paradigm , And satisfy Every data record in the data sheet , Are uniquely identifiable , And all non primary key fields , Must be completely dependent on the primary key , You can't rely on only part of the primary key .
• If you know the values of all attributes of the primary key , We can retrieve any tuple ( That's ok ) Any value of any property of ( The primary key in the requirement can be expanded and replaced with a candidate key )
• 

for instance , On the grade sheet ( Student number , Course no. , achievement ) In relationship ,( Student number , Course no. ) You can decide your grades , Because a student can take many courses , A course can also be selected by multiple students , Therefore, neither the student number nor the course number can independently determine the grade .
therefore ( Student number , Course no. )——> The result is Complete dependency .

The game table contains the player number , full name , Age , Game number , The nature of the game, the venue, etc , Both candidate keys and primary keys are ( Player number , Game number ), We can use candidate keys ( Primary key ) To determine the following relationship .
( Player number , Game number )——>( full name , Age , Time of the game , The venue , score )
But this data table does not satisfy the second paradigm , Because the fields in the data table still have the following correspondence :

( Player number )——>( full name , Age )
( Game number )——>( Time of the game , The venue )
Non primary attributes are not entirely dependent on candidate keys , This will cause the following problems .

• data redundancy : If a player participates m game , Then the player's name and age are repeated m-1 Time , A game may have n Two players participate in , The time and place of the game are repeated n-1 Time
• Insertion exception : If we want to add a new game , But it's not clear who the players are 2, Then there is no insert
• Delete exception : We want to delete a player number , But if you don't save the game table separately , Will delete the game information at the same time
• Update exception : If we adjust the time of a game , Then all the time of the game in the data sheet must be adjusted , Otherwise, there will be the same game but different times .

To avoid the above , We can design the players' game table into the following three tables .

Table name	attribute ( Field )
players player surface	Player number , full name , Age and other attributes
match game surface	Game number , Time of the game , Playing field and other attributes
The relationship between players and the game player_game surface	Player number , Game number , Score and other attributes

In this case , Each data sheet conforms to the second paradigm , To avoid the occurrence of abnormal conditions
The second paradigm requires that the attributes of an entity completely depend on the primary keyword , If there is an incomplete dependency , Then this attribute and this part of the main keyword should be separated to form a new entity , There is a one to many relationship between the new entity and the original entity

2.5 Third normal form (3NF）

• The third paradigm is based on the second paradigm
• Every non primary key field in the data table is directly related to the primary key field
• That is, all non primary key fields in the data table cannot depend on other non primary key fields
• This rule means that all non primary attributes cannot have dependencies , They are independent of each other
• The primary key here can be expanded into a candidate key
• 

2.6 The advantages and disadvantages of paradigms

• advantage :

  • The standardization of data helps to eliminate data redundancy in the database

• The third paradigm is generally considered to be in performance , The best balance between scalability and data integrity

• shortcoming :

  • It reduces the query efficiency , Because the higher the paradigm level , The more watches you design , When querying data, you may need to associate multiple tables , Not only is it expensive , And it may invalidate some indexes

  • The paradigm is only the standard of design , When actually designing , We may violate the principles of the paradigm for performance and read efficiency , Improve the read performance of the database by adding a small amount of redundant or duplicate data , Reduce associated queries , Realize the purpose of space for time

3. Anti normalization

3.1 summary

• Follow the principle of business priority
• First, meet the business needs , Then come in to reduce redundancy
• Sometimes we want to optimize query efficiency , Anti paradigm is also an optimization idea , We can improve the read performance of the database by adding redundant fields to the data table .
• 
  

3.2 New problems of anti paradigm

Although anti - normal form can exchange space for reality , Improve the efficiency of query , But anti paradigm also brings some new problems

• The storage space has become larger
• A field in a table has been modified , Redundant fields in another table should also be modified synchronously , Otherwise, the data will be inconsistent
• If you use stored procedures to support data updates , Delete and other operations , If the operation is frequent , It will consume system resources
• In the case of a small amount of data , The anti normal form does not embody the performance advantage , It may also complicate the design of the database .

3.3 Applicable scenarios of anti paradigm

When redundant information can Greatly improve query efficiency When , We will adopt anti paradigm optimization .

Suggestions for adding redundant fields
Adding redundant fields must meet the following two conditions , Add redundant fields only when the following two conditions are met
① This redundant field does not need to be modified frequently
② This redundant field is indispensable for query

4.BCNF( Bass paradigm )

This table conforms to the third normal form