SQL COUNT DISTINCT

Create table trade

(

　　sell_id int, --卖家

　　buy_id int, -- 卖家

time date --交易时间

)

sell_id, buy_id, time

s1, b2, 2013-1-1

s1, b3, 2013-1-1

s1, b2, 2013-1-2

s2, b4, 2013-1-2

select sell_id, count(buy_id) from trader group by sell_id

结果是： s1, 3

s2, 1

select count(distinct buy_id) from trader group by sell_id

结果是：s1, 2

s2, 1

总结： count distinct 来计算非重复结果的数目。