Summary of SQL query de duplication statistics methods

Speaking of sql When you go to a major home, you should think of distinct This keyword . But sometimes distinct Not suitable for some special scenes .

Examples of scenes :

The following figure shows my file storage table . Now there is a need to count the number of users who upload files .

 

If you use distinct The number of queries is shown in the figure below . The number is 8 Obviously not , The correct quantity should be 4, Repetitive user_id It shouldn't be counted for me .

distinct What if you can't meet this situation , Another common method is to use group by Group statistics . Please look directly at the code, which I wrote SELECT  COUNT(1) AS countUser FROM (SELECT COUNT(1)  FROM sys_file_info GROUP BY user_id ) AS a. This sentence is correct .

Thinking of grouping statistics , Let's first look at the grouping statistics . according to user_id Divide up 4 Group , The number of each group is 1,2,1,4. Then the number of groups is the number of users after de duplication , namely 4 Group =4 Users .

Let's count the number of groups at this time , How can we count the number of groups . It is very simple to use sub query , Take the table queried by grouping as a pseudo table and make statistics again .

SELECT  COUNT(1) AS countUser FROM (SELECT COUNT(1)  FROM sys_file_info GROUP BY user_id ) AS a  The red part is to form a pseudo table . The blue part is to count the number of entries in the pseudo table , Then we get the number of groups , That is, the number of users .