从 SQLite 中的 SQL 查询结果中删除重复项

Question

我已经编写了 SQL 查询并得到了以下结果，但我想根据每个产品的 min(sales) 从结果中删除重复的年份。

                SELECT petroleum_product AS Product,
                CAST(year / 5 * 5 AS CHAR) || '-' || 
                    CAST(year / 5 * 5 + 4 AS CHAR) AS Year,                       
                MIN(sale) AS MIN,      
                MAX(sale) AS Max,
                AVG(sale) AS AVG
                FROM REPORT
                GROUP BY Product, Year
                ORDER BY 2;

我从这个查询中得到了以下结果：

('Aviation Turbine Fuel', '2000-2004', 63131, 63131, 63131.0)  
('Aviation Turbine Fuel', '2000-2004', 47453, 47453, 47453.0)
('Aviation Turbine Fuel', '2000-2004', 52839, 52839, 52839.0)
('Aviation Turbine Fuel', '2000-2004', 64041, 64041, 64041.0)
('Aviation Turbine Fuel', '2000-2004', 66825, 66825, 66825.0)
('Diesel', '2000-2004', 326060, 326060, 326060.0)
('Diesel', '2000-2004', 286233, 286233, 286233.0)
('Diesel', '2000-2004', 299973, 299973, 299973.0)
('Diesel', '2000-2004', 299730, 299730, 299730.0)
('Diesel', '2000-2004', 315368, 315368, 315368.0)
('Aviation Turbine Fuel', '2010-2014', 101314, 101314, 101314.0)
('Aviation Turbine Fuel', '2010-2014', 109808, 109808, 109808.0)
('Aviation Turbine Fuel', '2010-2014', 115786, 115786, 115786.0)
('Aviation Turbine Fuel', '2010-2014', 123527, 123527, 123527.0)
('Aviation Turbine Fuel', '2010-2014', 139404, 139404, 139404.0)
('Diesel', '2010-2014', 655128, 655128, 655128.0)
('Diesel', '2010-2014', 648513, 648513, 648513.0)
('Diesel', '2010-2014', 716747, 716747, 716747.0)
('Diesel', '2010-2014', 811100, 811100, 811100.0)
('Diesel', '2010-2014', 901393, 901393, 901393.0)
('Aviation Turbine Fuel', '2005-2009', 64335, 64335, 64335.0)
('Aviation Turbine Fuel', '2005-2009', 63778, 63778, 63778.0)
('Aviation Turbine Fuel', '2005-2009', 68938, 68938, 68938.0)
('Aviation Turbine Fuel', '2005-2009', 68935, 68935, 68935.0)
('Aviation Turbine Fuel', '2005-2009', 82631, 82631, 82631.0)
('Diesel', '2005-2009', 294329, 294329, 294329.0)
('Diesel', '2005-2009', 306687, 306687, 306687.0)
('Diesel', '2005-2009', 302706, 302706, 302706.0)
('Diesel', '2005-2009', 446468, 446468, 446468.0)
('Diesel', '2005-2009', 612505, 612505, 612505.0)

经过我的查询，每年范围内的每个产品都有 5 个结果。但我希望最终结果有 min(sale) 是这 5 个中的最小值，max(sale) 是那些 5 中的最大值和 avg(sale) 是这 5 个中的平均值。其他年份范围的产品也是如此。

查询后的结果应该是这样的：

('Aviation Turbine Fuel', '2000-2004', 47453, 66825, 58857.8)
('Diesel', '2000-2004', 286233, 66825, 305472.8)
('Aviation Turbine Fuel', '2005-2009', 63778, 82631, 69723.4)
('Diesel', '2005-2009', 294329, 612505, 392539)
('Aviation Turbine Fuel', '2010-2014', 101314, 139404, 117967.8)
('Diesel', '2010-2014', 648513, 901393, 746576.2)

Answer 1

您可以使用 2 级聚合来做到这一点：

SELECT Product, 
       Year,
       MIN(min_sale) AS MIN,
       MAX(max_sale) AS MAX,
       AVG(avg_sale) AS AVG
FROM (
  SELECT petroleum_product AS Product,
         (year / 5 * 5) || '-' || (year / 5 * 5 + 4) AS Year,       
         MIN(sale) AS min_sale,      
         MAX(sale) AS max_sale,
         AVG(sale) AS avg_sale
  FROM REPORT
  GROUP BY Product, Year
)
GROUP BY Product, Year

在连接它们之前，无需将整数值转换为 CHAR。

Answer 2

如果我理解正确，您可以使用窗口函数：

select t.*
from (
    select 
        petroleum_product as product,
        cast(year / 5 * 5 as char) || '-' || cast(year / 5 * 5 + 4 as char) as year,                       
        min(sale) as min_sale,
        max(sale) as max_sale,
        avg(sale) as avg_sale,
        row_number() over(partition by product order by min(sale)) rn
    from report
    group by product, year
) t
where rn = 1
order by 2;

从您现有的查询开始，这将为每个 product 提供最少 min(sale) 的行。