添加出现次数的列数

Question

我在 SQL Server 中找到了名为 df 的表here:

-- Parameters
DECLARE @Year INT = 2020; --, @Country varchar(50)= 'Brazil';

WITH ModeData AS (
    SELECT country, 
           a.Mode
    FROM df
    CROSS APPLY (
        SELECT TOP 1 Mode, COUNT(*) AS cnt
        FROM (VALUES (val1), (val2), (val3)) AS t(Mode)
        GROUP BY Mode
        ORDER BY COUNT(*) DESC
    ) a
  where year=@year --and  country=@country 
)

-- Calculate proportions and map modes to labels
, Proportions AS (
    SELECT country, 
           CASE 
               WHEN Mode = 1 THEN 'Very Dissatisfied'
               WHEN Mode = 2 THEN 'Dissatisfied'
               WHEN Mode = 3 THEN 'Neutral'
               WHEN Mode = 4 THEN 'Satisfied'
               WHEN Mode = 5 THEN 'Very Satisfied'
           END AS SatisfactionLevel,
           COUNT(*) * 1.0 / SUM(COUNT(*)) OVER (PARTITION BY country) AS Proportion
    FROM ModeData
    GROUP BY country, Mode
)

-- Pivot the results to get each satisfaction level as a column
SELECT country, 
       [Very Dissatisfied], 
       [Dissatisfied], 
       [Neutral], 
       [Satisfied], 
       [Very Satisfied]
FROM Proportions
PIVOT (
    MAX(Proportion)
    FOR SatisfactionLevel IN ([Very Dissatisfied], [Dissatisfied], [Neutral], [Satisfied], [Very Satisfied])
) AS p
ORDER BY country;

结果表是：

国家	非常不满意	不满意	中性	满意	非常满意
巴西	0.285714285714	0.142857142857	0.142857142857	0.142857142857	0.285714285714
加拿大	0.111111111111	0.111111111111	0.333333333333	0.222222222222	0.222222222222
法国	0.250000000000	0.125000000000	0.250000000000	0.250000000000	0.125000000000
意大利	0.166666666666	0.166666666666	0.166666666666	0.166666666666	0.333333333333
美国	0.222222222222	0.111111111111	0.111111111111	0.333333333333	0.222222222222

我想计算每个国家/地区的计数。表 df 中每个国家/地区有多少行，并将此计数添加为结果表中的额外列。理想情况下，基于玩具示例数据，我想要的结果如下：

国家	非常不满意	不满意	中性	满意	非常满意	数
巴西	0.285714285714	0.142857142857	0.142857142857	0.142857142857	0.285714285714	7
加拿大	0.111111111111	0.111111111111	0.333333333333	0.222222222222	0.222222222222	9
法国	0.250000000000	0.125000000000	0.250000000000	0.250000000000	0.125000000000	8
意大利	0.166666666666	0.166666666666	0.166666666666	0.166666666666	0.333333333333	6
美国	0.222222222222	0.111111111111	0.111111111111	0.333333333333	0.222222222222	9

Answer 1

添加额外的 cte 来解决国家/地区计数，然后将其包含在最终结果中。

DECLARE @Year INT = 2020; --, @Country varchar(50)= 'Brazil';

WITH ModeData AS (
    SELECT country, 
           a.Mode
    FROM df
    CROSS APPLY (
        SELECT TOP 1 Mode, COUNT(*) AS cnt
        FROM (VALUES (val1), (val2), (val3)) AS t(Mode)
        GROUP BY Mode
        ORDER BY COUNT(*) DESC
    ) a
  where year=@year --and  country=@country 
)

-- Calculate proportions and map modes to labels
, Proportions AS (
    SELECT country, 
           CASE 
               WHEN Mode = 1 THEN 'Very Dissatisfied'
               WHEN Mode = 2 THEN 'Dissatisfied'
               WHEN Mode = 3 THEN 'Neutral'
               WHEN Mode = 4 THEN 'Satisfied'
               WHEN Mode = 5 THEN 'Very Satisfied'
           END AS SatisfactionLevel,
           COUNT(*) * 1.0 / SUM(COUNT(*)) OVER (PARTITION BY country) AS Proportion
    FROM ModeData
    GROUP BY country, Mode
)

, Pivoted AS (
-- Pivot the results to get each satisfaction level as a column
SELECT Country, 
       [Very Dissatisfied], 
       [Dissatisfied], 
       [Neutral], 
       [Satisfied], 
       [Very Satisfied]
FROM Proportions
PIVOT (
    MAX(Proportion)
    FOR SatisfactionLevel IN ([Very Dissatisfied], [Dissatisfied], [Neutral], [Satisfied], [Very Satisfied])
) AS p
),

CountryCounts AS (
-- Count of countries from original df table
SELECT Country,
       COUNT(Country) AS Total
FROM df
GROUP BY Country),

SELECT Pivoted.Country,
    [Very Dissatisfied],
    Dissatified,
    Neutral,
    Satisfied,
    [Very Satisified],
    Total
FROM Pivoted
INNER JOIN CountryCounts
ON Pivoted.Country = CountryCounts.Country
ORDER BY Pivoted.Country;