Notes on MySQL related knowledge points (startup, index)

linux Service startup under

service mysqld start

systemctl start mysqld.service   #（ The above command is actually equivalent to executing the following command 【 Will jump 】）

Connect mysql The server

mysql  -uroot -p Input password

Operating the database

create database  Database name    # Create database 

drop database  Database name   # Delete database 

use  Database name   # Select database 

Index related

mysql The indexes of all seem to be based on B+ Tree built , unique index 、 Primary key 、 General index 、 Full text index, etc .

Logical classification

General index

#  Create index 
create index indexName on tableName(columnName(length));

#  Modify table structure to add index 
alter table tableName add index indexName (columnName(length));

#  Create index when creating table 
create table tableName(
id int primary key,
index indexName (columnName(10))  # Create index 
);

#  Delete index < General index deletion method >
#  Delete an index of a table 
drop index indexName on tableName;
#  Delete an index of a table 
alter table tableName drop index indexName;

unique index

#  Create unique index 
create unique index indexName on tableName(columnName(length));

#  Modify the table to create a unique index 
alter table tableName add unique index indexName (columnName(length));

# Add a unique index when creating a table 
create table tableName(
    id int primary key,
    name varchar(10),
    unique index indexName(name(length))
);

Primary key

#  Create primary key 
create table tableName(
id int primary key auto_increment
);
#  perhaps 
create table tableName(
id int auto_increment,
primary key (id);
);

#  Set the primary key separately 
alter table tableName add primary key (id);
 
#  Delete index 
#  If the primary key is self increasing , We need to get rid of self increase 
alter  table tableName modify id int ;    #  Remove autoincrement 
alter table tableName drop primary key;

Full-text index

View index information

show index from tableName;

Physical classification

Clustered index

It is generally a primary key index , If there is no primary key index , Just choose a non empty unique index , If not , Just choose rowid As a clustered index .

Each table has only one clustered index .

A clustered index is actually the storage structure of a table , It's a tree structure . If there is no clustered index , Data is neatly arranged and stored on disk .

  Nonclustered indexes

The logical order of the index is generally different from the storage order of the data , A table can have multiple nonclustered indexes . Leaf nodes in nonclustered indexes do not store actual data , Save only the reference of the clustered index （ Such as primary key id）.

Each table can have multiple nonclustered indexes .

  Index advantages and disadvantages

advantage

Through the index , Reduce io frequency , Improve query efficiency

shortcoming

1、 Create index , Take up storage space

2、 When adding, deleting, or modifying an index , To maintain the structure of the index , Affect efficiency

Indexing rules

Should be created

1、 Have a primary key

2、 Frequently as queries 、 Fields for grouping and sorting criteria

3、 Fields associated with other tables

Avoid creating

1、 Indexing is not recommended for those with high repetition values , The amount of data returned to the table for query is too large （ Second inquiry ）

2、 Do not index frequently updated fields , Because the maintenance cost is too high

3、 Tables with a small amount of data do not need to be indexed

Index usage considerations

1、like sentence '% Zhang San ' , The index will fail

2、 Unconditional query , Index failure , Such as   != 、<> 、not in 、not like wait

3、 Joint index , Pay attention to order , The conditions with strong screening should be ranked first （ Highly unique ）

4、 Cast does not go through the index .

5、 Use overlay index to avoid the occurrence of back table , Improve query efficiency

Overlay index is to change a single column non primary key index into a multi field joint index , You can find the desired data in an index tree , There is no need to go to the primary key index tree , Search again This phenomenon , be called Index overlay .

See ---MySQL Coverage index details - Nuggets This is a post study experience sharing , Before body , We need to explain some basic concepts of index . 1. What is index ? Indexes （ stay MySQL It's also called “ key key”） It's a data structure that the storage engine quickly finds records , Generally speaking, it is similar to the catalogue of books , Although this analogy is used most, it is also the most appropriate , Do not borrow a certain knowledge point in the query book …https://juejin.cn/post/6844903967365791752

6、null 、not null 、!= You may also use the index , When the amount of returned data is less than 20% I'll go .

See ---

MySQL in IS NULL、IS NOT NULL、!= No index ？ Rats ！ - Nuggets It's getting worse , It is even regarded as the truth by many students . We don't say anything , for instance . If we had a watch s1, The structure is as follows ： There are several query statements above WHERE Used in Clause IS NULL、IS NOT NULL、!= These conditions , But it can be seen from their implementation plans , These statements use the corresponding secondary index to execute the query , Instead of using the so-called …https://juejin.cn/post/6844903921450745863

7、 Optimize paging scenarios with sub queries

MySQL It's not about skipping offset That's ok , It's about taking offset+N That's ok , Before giving up offset That's ok , return N That's ok , That's right offset When I was very old , Very inefficient , Or control the total number of returns , Or for pages that exceed a certain threshold SQL rewrite .

 8、 Business unique fields , Even a combination of multiple fields , You should also create indexes （ Composite index ）

 9、 When establishing Association for multi table query , Ensure that the corresponding fields have indexes and the data types on both sides are consistent

10、 If the result is clear, there is only one , Use limit = 1 Can improve efficiency

11、select Statement declaration field

12、 As far as possible with union all Instead of union.Union The collection needs to be merged before uniqueness filtering , This involves sorting , a large number of cpu operation , Increase resource consumption and delay , Of course , Use union all The precondition is that there is no duplicate data in the two result sets .