问题
I have the following medleys table that combines colors, fruits and ratings:
[medleys]
medley_id | color | fruit | rating
==============================================
1 red apple 25
2 blue pear 5
3 green apple 12
4 red apple 10
5 purple kiwi 5
6 purple kiwi 50
7 blue kiwi 3
8 blue pear 9
I am trying to write an ANSI-compliant SQL query that will combine every unique/distinct color-fruit pair and sum each pair's individual rating values. Thus if you ran the query on the table above it would produce the following result sets:
[query]
color | fruit | sum
===========================
red apple 35
blue pear 14
blue kiwi 3
green apple 12
purple kiwi 55
Thus, the query sees there are two red-apple pairs in the table, and so it creates one result for the red-apple pair, and adds up their constituent ratings (25 + 10 = 35), etc.
I am sure that I need to do a select for distinct color/fruit values, but not sure how to aggregate the ratings at the same "level/scope":
SELECT
distinct(color, fruit), sum(rating)
FROM
medleys
Order doesn't matter. color and fruit are VARCHAR(50)s and rating is INT. Thanks in advance!
回答1:
SELECT color, fruit, sum(rating)
FROM medleys
GROUP BY color, fruit
Distinct is used to select distinct elements, nothing more, while you want to aggregate and for that you need GROUP BY and aggregation functions (SUM).
回答2:
You don't need distinct at all. You need group by:
select color, fruit, sum(rating)
from medleys
group by color, fruit
I'm answering, because I see this mistake occur. In general, you don't need select distinct at all in SQL. You can always use a group by instead. Distinct should be introduced after group by as a convenient short-hand.
回答3:
SELECT `color`,`fruit`,SUM(`rating`)
FROM Medleys
GROUP BY `color`,`fruit`
SQL Fiddle Example
回答4:
This should answer your question:
SELECT color, fruit, sum(rating) as [sum]
FROM medleys
GROUP BY color, fruit
ORDER BY color
来源：https://stackoverflow.com/questions/15512628/sql-query-with-distinct-and-sum