Reading mysql45 lecture - index continued

There is a table creation statement as follows ：

 create table T ( ID int primary key, k int NOT NULL DEFAULT 0, s varchar(16) NOT NULL DEFAULT '', index k(k)) engine=InnoDB;
insert into T values(100,1, 'aa'),(200,2,'bb'),(300,3,'cc'),(500,5,'ee'),(600,6,'ff')

In execution select * from t where k between 3 and 5 The general process of ：

1. First, according to k Index tree found k=3 Index entry for
2. And then depending on the corresponding ID value =300 Query the whole row of data back to the table
3. And then k The index tree continues to look for the next value , Get next k=5 Index entry for
4. obtain k=5 Corresponding ID=500, Query the whole row of data in the table again
5. Continue to k Search the index tree for the next value , Find the next k=6 It doesn't meet the requirements , So stop querying , And the operation of returning to the table will not be performed .

As mentioned before ： The data in the table is stored according to the data of the primary key , So when you need to query the data in the table, you must return to the table , So is there any situation that you don't need to go back to the table ？

Overlay index

If the above query statement is select ID from t where k between 3 and 5 , That is, you only need to query ID Value , You don't need a whole row of data , and ID The value of already exists in K On the index tree , So there is no need to return the form at this time .

Tree searches can be reduced by overwriting the index , Significantly improve query performance , So using overlay index is a common By means of performance optimization .

Suppose there is a citizen list ：

CREATE TABLE `tuser` ( 
`id` int(11) NOT NULL, 
`id_card` varchar(32) DEFAULT NULL,
 `name` varchar(32) DEFAULT NULL,
  `age` int(11) DEFAULT NULL, 
  `ismale` tinyint(1) DEFAULT NULL, 
  PRIMARY KEY (`id`),
   KEY `id_card` (`id_card`),
   KEY `name_age`  (`name`,`age`) ) ENGINE=InnoDB

Because the ID number can identify a person , Therefore, it is often necessary to use the ID number to query personal information .

Suppose there is a high frequency request , Is to use the ID card number to query the name ;
that sql Almost select name from tuser where id_card = xxxx; because name Not the primary key. , No existence id_card On the index tree , You need to go back to the table ;
In order to reduce the number of back tables , We will name and id Being a federated primary key is a little far fetched , Because it is impossible to use the fields required by each high-frequency request as the joint primary key .

So this time you can id_card and name Create a joint index ;

Index push down

Go one step further on the request for names , The query name contains the name of the table , The ID number is 32 start ;
select name from tuser where substring(id_card,1,2) = 32 and name like ‘% surface %’

stay mysql5.6 Before , On the basis of substring(id_card,1,2) = 32 Find eligible ID after , Will return to the table again , Inquire about name Field and then determine whether the word table is included ;

stay mysql5.6 after , On the basis of substring(id_card,1,2) = 32 Find eligible ID after , It will directly judge name Whether the field contains the word table ;

Leftmost prefix principle

Based on the results of the above joint index , Then, if you still check the vaccination status according to your ID card number , Check the loan times and other background according to the ID card number , In order to reduce the operation of returning tables , Do we all have a joint index , There are too many indexes ;
You can see that there is already a (name,age) The index of ：

You can see that the index entries are sorted according to the field order defined by the index .
It's just (name,age) This joint index is name before ,age After that ; So the index entries are all name before ,age After ; Index entries are based on name Sort , And then based on age Sort , because name= Zhang San has more than one , So we will follow the internal rules age Sort ;

So when we execute select Number of loans from table where name like '% Spongebob %' When , Also can walk (name,age) This index ;
But if it is execution select Number of loans from table where age >40 I won't go (name,age) This index ;
This is the leftmost prefix principle , As long as the leftmost prefix principle is satisfied , You can use indexes to speed up queries .

The joint index is (a,b); But there are only b The limitations of , You won't use the union index , If you want to go to the index at this time , You have to maintain a separate (b) The index of ;
But if the union index is (b,a), There are only... In the query criteria b The limitations of , That is to say, they will query , You can maintain one index less , Save space .

If by adjusting the order , can The first Maintain one index less , So this order is often the one that needs to be prioritized .