Question

有没有其他方法可以编写此查询？我尝试在子查询中执行它，但由于多列而无法工作。看起来这个查询只能单独使用。请纠正我

记录

PK recordId

dateViewed

CarViewed

我试过这个

   SELECT R.`dateViewed` FROM Records ,(
   SELECT R.CarViewed, COUNT(R.CarViewed) as cnt FROM  Records R
   GROUP BY R.CarViewed
   ORDER BY cnt DESC
   LIMIT 1 ) AS favouriteCarOfTheDay
   GROUP BY R.`dateViewed

然后我尝试了这个

   SELECT R.`dateViewed` ,COUNT(R.CarViewed) as cnt FROM Records ,
   (
   SELECT R.CarViewed FROM  Records R
   GROUP BY R.CarViewed
   ORDER BY cnt DESC
   LIMIT 1 ) AS favouriteCarOfTheDay
   GROUP BY R.`dateViewed

在我尝试的许多其他查询中，我不知道如何让它工作。

简而言之，我希望得到最常见的汽车。

喜欢： dateViewed favouriteCarOfTheDay

2012-09-22 | Nissan

2012-09-23 | BMW

Answer 1

试试这个

   SELECT R.`dateViewed` ,COUNT(R.CarViewed) as cnt ,R.CarViewed FROM Records R
   GROUP BY R.`dateViewed 
   ORDER BY COUNT(R.CarViewed) DESC

Answer 2

我认为以下内容应该有效（免责声明，而不是我自己的所有工作，改编自another question的答案）

SELECT DISTINCT
    R.dateViewed,
    R.CarViewed
FROM Records R
WHERE
    R.dateViewed =
        (SELECT R2.dateViewed FROM
            (
                SELECT R1.dateViewed, COUNT(*) AS numViewed
                FROM Records R1 
                WHERE R1.CarViewed = R.CarViewed
                GROUP BY R1.dateViewed
                ORDER BY R1.numViewed DESC
            ) AS R2
            LIMIT 1
        )
ORDER BY r.dateViewed

Answer 3

这样的事情在MySQL中真的很糟糕，所以它实际上可能比两个相关的子查询慢，但至少它会返回汽车和它的viewcount：

SELECT counts.`dateViewed`,counts.`CarViewed` as favourite_car, counts.cnt
FROM
(SELECT R.`dateViewed` ,R.`CarViewed`, COUNT(*) as cnt 
FROM Records
GROUP BY R.`dateViewed` ,R.`CarViewed`
) as counts JOIN
(SELECT R.`dateViewed`, MAX(cnt) as cnt
FROM
(SELECT R.`dateViewed` ,R.`CarViewed`, COUNT(*) as cnt 
FROM Records
GROUP BY R.`dateViewed` ,R.`CarViewed`
) as q
GROUP BY R.`dateViewed`) as maxes
ON counts.cnt=maxes.cnt

最常见的SubQuery MySql

3 个答案: