[MySQL learning notes 25] SQL statement optimization

insert Optimize

1. Multiple single insert Statements can be merged into a batch insert insert sentence . One time insertion is better than many times , Because each insertion requires a connection .
2. Manually control transactions . because mysql The transaction of is automatic by default , Every time you execute one sql Statement will start a transaction .
3. Insert primary key in order . This sum mysql Bottom related .
4. Mass data import is not recommended insert, It is load Instructions .

# When the client connects to the server , Add parameters –local-infile
mysql --local-infile -u root -p
# Set global parameters local_infile by 1, Turn on the switch to import data from the local load file
set global local_infile = 1;
# perform load The instruction will prepare the data , Load into table structure
load data local infile ‘/root/sql1.log’ into table ‘tb_user’ fields terminated by ‘,’ lines terminated by ‘\n’;

Primary key optimization

1. stay innodb In the engine , Table data is organized and stored according to the primary key , This storage method is called index organization table (index organized table IOT)
When the primary key is inserted out of order, it is easy to split pages , Therefore, disordered insertion should be avoided , Or you can directly use the auto increment primary key , Once and for all
expand ： Page merge , When the amount of data in a page is less than 50% when ,mysql Will try to merge adjacent pages , Called page merge
2. The length of the primary key should be minimized in development , Because secondary indexes store primary keys , The longer it takes up more space

order by Optimize

using filesort： Put the data into the cache , Sort .
using index： Directly use the index to return the sorting results .
mysql Due to the use of B+Tree Indexes , It is easy to traverse the data in order , Index when order by When the following field has an index , No additional sorting is required , bring order by Increased efficiency .
If the overlay index is not used , Then it is unrealistic to use index sorting , Because the field is missing , So we need to do filesort.
filesort The buffer size of is limited , If the data exceeds the buffer, it will be sorted on the disk , Greatly affect performance .
The buffer size can be adjusted .

group by Optimize

using temporary： Use temporary tables to group , Low efficiency
using index： Use index grouping , Efficient

Set table user Yes (profession,age,status) Joint index
select profession from user group by profession;#using index
select age,count(age) from user group by age;#using index;using temporary
notes ： There's one in front using index Because the overlay index
select profession , age, count() from tb_user group by profession,age;#using index
select age, count() from tb_user where profession = ‘ Software Engineering ’ group by age;#using index

limit Optimize

select * from tb_sku limit 2000000,10;
Replace with
select * from tb_sku t , (select id from tb_sku order by id limit 2000000,10) a where t.id = a.id;

count Optimize

myisam The engine uses a separate variable to record the total number of records in the table , therefore count It's very efficient , If not where
innodb The engine needs a full table scan , To count the quantity
Optimization idea ： Open an extra place to store the total quantity
count( Field )<count( Primary key )<count(1)≈count(*)

update Optimize

When updating, try to update according to the indexed fields , This is a row lock , Otherwise, it will become a table lock .