Home  >  Article  >  Database  >  Solving SQL problems will definitely take your understanding of MySQL one step further!

Solving SQL problems will definitely take your understanding of MySQL one step further!

coldplay.xixi
coldplay.xixiforward
2020-12-26 17:42:094821browse

SQL Tutorial This column introduces how to understand MySQL more effectively

Solving SQL problems will definitely take your understanding of MySQL one step further!

##Recommended (free): SQL tutorial

The attribute table (product_props) structure is as follows

Data volume is more than 800W

Field nameTypeDescriptionidintidpn_idintAttribute typepv_idintAttribute valueproduct_idintProduct ID
There is a one-to-many relationship between product_id, pn_id, and pv_id.

The data is similar to this:

product_idpn_idpv_id109705 (Model) 135 (Apple 9) 1097011 (Memory) 23 (512G)1097010 (Color)17 (Local gold)109708 (Network)6(5G)109805135109801124 (1024G) ##10980
10 16 (Aurora Blue)
The product table (product) structure is as follows

data Amount above 40W

Field nameproduct_idtype_idbrand_idmodel_idstatusThe data is similar to the following:
Type Description
int product_id
int typeid
int brandid
int Model id
tinyint status

product_id109701098010981
type_id brand_id model_id status
1(Mobile) 1(Apple) 1(Iphone8) 1(Normal)
1(Mobile) 1(Apple) 1(Iphone8X) 3 (Sold)
1(Mobile) 1(Apple) 1(Iphone8XP) 1(Normal)
Problem

Find out the model number is

Apple 9

At the same time, the memory is 512G, color is local gold, status is normal, the total number of products is , ps: attribute conditions may exceed 10 Group.
Requirements

Performance comes first, avoid aggregate functions, etc.

Performance ranking of solutions to the original problem

Exist scheme from @Kamicloud
  1. SELECT 
        sql_no_cache `product_id`
    FROM
        `zx_tests` AS a
    WHERE
        `pn_id` = 101 AND `pv_id` = 59
            AND EXISTS( SELECT 
               sql_no_cache  *
            FROM
                `zx_tests`
            WHERE
            a.product_id = product_id and
                `pn_id` = 101 AND `pv_id` = 171);
    
        2 组条件下 0.657,3 组 0.695,4 组 0.759,5 组 0.743 (单独查属性表)
Subquery scheme from @Elijah_Wang
  1. SELECT `product_id` FROM `product` WHERE `pn_id` = 5 AND `pv_id` = 135 AND `product_id` IN (SELECT `product_id` FROM `product` WHERE `pn_id` = 11 AND `pv_id` = 23);
    
        2 组条件下 0.729,3 组 0.75,4 组 0.730,5 组 0.757 (新问题之前)
Performance ranking after new questions

Subquery plan from @Elijah_Wang
  1.     select SQL_NO_CACHE count(1) from pdi_product a join  (
            SELECT
                 distinct product_id
            FROM
                `product_props` 
            WHERE
                `pn_id` = 5 
                AND `pv_id` = 127
                AND `product_id` IN ( SELECT  `product_id` FROM `product_props` WHERE `pn_id` = 11 AND `pv_id` = 22 ) 
                AND `product_id` IN ( SELECT  `product_id` FROM `product_props` WHERE `pn_id` = 10 AND `pv_id` = 18 ) 
                AND `product_id` IN ( SELECT  `product_id` FROM `product_props` WHERE `pn_id` = 8 AND `pv_id` = 6 )  
                AND `product_id` IN ( SELECT  `product_id` FROM `product_props` WHERE `pn_id` = 9 AND `pv_id` = 1 )
                ) b on a.product_id = b.product_id 
                where  a.status = 1;
  2. It takes 1.5-1.56 (range of 10 executions)

expain analysis:

        select SQL_NO_CACHE count(1) from pdi_product a 
            where  a.status = 1 and a.product_id in (SELECT
             distinct product_id
        FROM
            `product_props` 
        WHERE
            `pn_id` = 5 
            AND `pv_id` = 127
            AND `product_id` IN ( SELECT  `product_id` FROM `product_props` WHERE `pn_id` = 11 AND `pv_id` = 22 ) 
            AND `product_id` IN ( SELECT  `product_id` FROM `product_props` WHERE `pn_id` = 10 AND `pv_id` = 18 ) 
            AND `product_id` IN ( SELECT  `product_id` FROM `product_props` WHERE `pn_id` = 8 AND `pv_id` = 6 )  
            AND `product_id` IN ( SELECT  `product_id` FROM `product_props` WHERE `pn_id` = 9 AND `pv_id` = 1 ))
SQL 性能实战来了,机不可失! takes 0.69-0.72 (range of execution 10 times)

explain analysis:

  • SQL 性能实战来了,机不可失!
exist solution from @Kamicloud
  1.     SELECT SQL_NO_CACHE
                count(1) 
            FROM
                product a
            WHERE
                a.STATUS = 1 
                AND a.product_id IN (
            SELECT DISTINCT
                `product_id` 
            FROM
                `product_props` AS a 
            WHERE
                a.`pn_id` = 5 
                AND a.`pv_id` = 127 
                AND EXISTS ( SELECT product_id FROM `product_props` WHERE a.product_id = product_id AND `pn_id` = 11 AND `pv_id` = 22 ) 
                AND EXISTS ( SELECT product_id FROM `product_props` WHERE a.product_id = product_id AND `pn_id` = 10 AND `pv_id` = 18 ) 
                AND EXISTS ( SELECT product_id FROM `product_props` WHERE a.product_id = product_id AND `pn_id` = 9 AND `pv_id` = 1 ) 
                AND EXISTS ( SELECT product_id FROM `product_props` WHERE a.product_id = product_id AND `pn_id` = 8 AND `pv_id` = 6 ) 
                );
  2. It takes 5.7-5.85 (range of execution 10 times)

explain analysis:

    SELECT SQL_NO_CACHE
        count(1) 
    FROM
        pdi_product a
        join (SELECT DISTINCT
        `product_id` 
    FROM
        `product_props` AS a 
    WHERE
        a.`pn_id` = 5 
        AND a.`pv_id` = 127 
        AND EXISTS ( SELECT product_id FROM `product_props` WHERE a.product_id = product_id AND `pn_id` = 11 AND `pv_id` = 22 ) 
        AND EXISTS ( SELECT product_id FROM `product_props` WHERE a.product_id = product_id AND `pn_id` = 10 AND `pv_id` = 18 ) 
        AND EXISTS ( SELECT product_id FROM `product_props` WHERE a.product_id = product_id AND `pn_id` = 9 AND `pv_id` = 1 ) 
        AND EXISTS ( SELECT product_id FROM `product_props` WHERE a.product_id = product_id AND `pn_id` = 8 AND `pv_id` = 6 ) ) b
    on a.product_id = b.product_id
    WHERE
        a.STATUS = 1
SQL 性能实战来了,机不可失!It takes 5.7-6.0 (range of execution 10 times)

explain analysis:

SQL 性能实战来了,机不可失!

#It can be seen that if you simply check the attribute table, the speed of the first position is the fastest, but when you check the product status, the speed is not as good as the subquery.

After explain analysis, the reason why the first subquery is fast is because its SQL is simple and the select_type is simple.

Regardless of join or exists method, select_type is mostly DERIVED and DEPENDENT SUBQUERY.

Related free learning recommendations:

mysql video tutorial

The above is the detailed content of Solving SQL problems will definitely take your understanding of MySQL one step further!. For more information, please follow other related articles on the PHP Chinese website!

Statement:
This article is reproduced at:learnku.com. If there is any infringement, please contact admin@php.cn delete