每年在SQL中查找最頻繁的數據

Question

表A：

Years Data
2000  A
2000  B
2000  C
2000  C
2000  D
2001  A
2001  B
2001  B
2002  B
2002  D
2002  D

我想輸出：

Years Data
2000  C
2001  B
2002  D

我的解決方案：

SELECT DISTINCT Years, Data
FROM
(
   SELECT Years, Data, COUNT(*) AS _count
   FROM TableA
   GROUP BY Years, Data
) a1
ORDER BY Years, _count DESC

但這有一個問題：

如果指定了SELECT DISTINCT則ORDER BY項目必須出現在選擇列表中。

如何更正我的SQL代碼？

Answer 1

假設您的數據庫支持row_number（），則可以這樣進行：

SELECT Years, Data
FROM 
(
    SELECT  Years,
            Data,
            ROW_NUMBER() OVER(PARTITION BY Years ORDER BY count(*) DESC) rn

    FROM TableA
    GROUP BY Years, Data
) x
WHERE rn = 1
ORDER BY Years, Data

觀看rextester上的現場演示。

Answer 2

嘗試這個：

select t.Years, t.[Data]
from (
    select *, count(*) cnt
    from TableA
    group by years, [Data]
    ) t
left join (
    select Years, max(cnt) maxCnt
    from (
        select *, count(*) cnt
        from TableA
        group by years, [Data]
    ) t
    group by Years
    ) tt on t.Years = tt.Years   -- tt is a view that gives you max count of each year
where t.cnt = tt.maxCnt          -- you need `years`, `[Data]` that its count is max count
order by years;

SQL小提琴演示

另一種方法是在支持它的DBMS中使用rank() ：

;with t as (
    select *, count(*) cnt
    from TableA
    group by years, [Data]
), tt as (
    select *, rank() over (partition by years order by cnt desc) rn
    from t
)
select years, [Data]
from tt
where rn = 1
order by years;

SQL小提琴演示

Answer 3

如果使用的是oracle，則可以使用STATS_MODE函數

  select years, stats_mode(data)
    from tablet
group by years;

https://docs.oracle.com/cd/B19306_01/server.102/b14200/functions154.htm

Answer 4

您的錯誤是"ORDER BY items must appear in the select list if SELECT DISTINCT is specified."

這意味着您已在ORDER BY中放入了不在SELECT 。 在這種情況下， _count DESC不在SELECT語句中

SELECT DISTINCT Years, Data, _count DESC
FROM
(
   SELECT Years, Data, COUNT(*) AS _count
   FROM TableA
   GROUP BY Years, Data
) a1
ORDER BY Years, _count DESC