mySQL Count the number of different values in a column
This requirement is actually very common. For example, we have a user source table to mark the channel from which users signed up. The table structure is as follows...
where origin is the source of users, and the values are iPhone, Android, and Web. Now it is necessary to count the number of users registered by these three channels.
Solution 1
SELECT count(*)
FROM user_operation_log
WHERE origin = 'iPhone';
SELECT count(*)
FROM user_operation_log
WHERE origin = 'Android';
SELECT count(*)
FROM user_operation_log
WHERE origin = 'Web'; count
separately with where statement their respective quantities.
In this way, the amount of query is a bit too much. If there are 10 values, then 10 similar statements have to be written, which is very troublesome.
Is there a single sentence to do it? So I did some research.
Solution 2
We know that count can be used to count not only the number of rows, but also the number of column values, for example:
Count how many rows there are in user_operation_log:
SELECT count(*) FROM user_operation_log
Count the number of origin columns whose value is not NULL:
SELECT count(origin) FROM user_operation_log
So we can use this feature to achieve the above requirements
The first way of writing (implemented with count) ）
SELECT
count(origin = 'iPhone' OR NULL) AS iPhone,
count(origin = 'Android' OR NULL) AS Android,
count(origin = 'Web' OR NULL) AS Web
FROM user_operation_log; The second way of writing the
query result
search_result
(implemented with sum)
SELECT
sum(if(origin = 'iPhone', 1, 0)) AS iPhone,
sum(if(origin = 'Android', 1, 0)) AS Android,
sum(if(origin = 'Web ', 1, 0)) AS Web
FROM user_operation_log;
SELECT
sum(origin = 'iPhone') AS iPhone,
sum(origin = 'Android') AS Android,
sum(origin = 'Web') AS Web
FROM user_operation_log;
第4种，最优的：
SELECT origin,count(*) num FROM user_operation_log GROUP BY origin;