使用GROUP BY有条件，怎么样？

Question

我想我可以如何解决我的任务：

SELECT row, val1, val2, SUM(1) AS total
FROM mess
if(row < 0){ GROUP BY val1, val2 } if(row > 0) {GROUP BY val1} 

我需要在一个SELECT中使用两个不同的GROUP BY，它是否可能？

Answer 1

SELECT row, val1, val2, SUM(1) AS total, 
    CASE WHEN row<0 THEN val2 ELSE 0 END as group2
FROM mess
GROUP BY val1, group2

如果row＆gt; 0，则所有值都相同，因此group result将仅计算val1

Answer 2

试试这个：

SELECT row, val1, val2, SUM(1) AS total
FROM mess
GROUP BY if(row>=0,val1,val1+','+val2)