How to analyze fans' interests?

【 subject 】

There is a “ Fan attention table ”, contain 3 A field ： user id、 Follow the media id、 date .

【 problem 】“ Fan attention table ” There is a situation that one user pays attention to multiple media at the same time , such as ： user id by A001 Users of , Focus on the media id The data is 1010,1020,1031. In order to facilitate the later analysis of fans' interests , Please split this situation in the table into multiple .

For example, for users A001, Its conversion is as follows ：

【 Their thinking 】

Such problems are called “ Column turned ”, stay MySQL There are generally three steps to deal with it ：

1） Create a “ Sequence table ”;

2） Join multiple tables , Copy each piece of data in the original table into multiple pieces ;

3） Use substring_index Function to get the final result .

First step ： Create sequence table

“ Sequence table ” It means that there is only one field , Stored is a sequence of numbers , such as ：

among ,“ Sequence ” The maximum value of is the maximum number of media a user pays attention to in this problem .

select max(length( Follow the media id) - length(replace( Follow the media id,',','')) + 1) as  The maximum number of media attention 
from  Fan attention table ;

Copy

The return result is ：

Then we need new “ Sequence table ” Namely ：

The second step ： Multi table join

Use multi table join , Can pass “ Sequence table ” take “ Fan attention table ” Each line of becomes multiple lines .

Here are two points to note ：

1） To ensure that every piece of data in the original table is not lost , choice “ Left link ”, And take the original table as the left table ;

2） The number of copies is limited in the connection condition , The limiting condition is the number of media users pay attention to , namely “ Follow the media id” The number of commas under the field plus 1.

select t1. user id,
       t1. Follow the media id,
       t1. date ,
       t2. Sequence 
from  Fan attention table  t1
left join  Sequence table  t2 on t2. Sequence  <= (length( Follow the media id) - length(replace( Follow the media id,',','')) + 1);

Copy

The return result is ：

The third step ： Use the function to get the result

The next step is to put the media id Intercept it , You need to use the string interception function ：SUBSTRING_INDEX.

SUBSTRING_INDEX( character string , Separator , Parameters )

Copy

among , Separator refers to dividing media in this question id Of “,”;2 Means to separate by separator , Intercept several media from left to right id; If the parameter is negative , It means to intercept several media from right to left id.

select t1. user id,
       substring_index(substring_index(t1. Follow the media id,',',t2. Sequence ),',',-1) as  Follow the media id,
       t1. date 
from  Fan attention table  t1
left join  Sequence table  t2 on t2. Sequence  <= (length( Follow the media id) - length(replace( Follow the media id,',','')) + 1);

Copy

The return result is ：

【 The test point of this question 】

1） Check your understanding of the ordered list ;

2） Check the string interception function SUBSTRING_INDEX Understanding ;

3） Check your understanding of multi table connections .