在 SQL 中组合类似的行并聚合一些值答案

【问题标题】：Combining like rows in SQL and aggregating some values在 SQL 中组合类似的行并聚合一些值
【发布时间】：2019-08-30 23:24:30
【问题描述】：

我有一些无意义数据的示例表，显示如下：

+--------+---------+---------+--------+---------+
| Vendor |   Day   |  Item   | Amount | Revenue |
+--------+---------+---------+--------+---------+
| Bob    | Monday  | Apple   |      1 |    1.00 |
| Bob    | Monday  | Orange  |      1 |    1.00 |
| Bob    | Monday  | Peach   |      2 |    2.00 |
| Bob    | Monday  | Lettuce |      3 |    3.00 |
| Bob    | Monday  | Spinach |      5 |    5.00 |
| Bob    | Monday  | Cabbage |      3 |    3.00 |
| Bob    | Tuesday | Apple   |      2 |    2.00 |
| Bob    | Tuesday | Orange  |      2 |    2.00 |
| Bob    | Tuesday | Peach   |      0 |    0.00 |
| Bob    | Tuesday | Lettuce |      3 |    3.00 |
| Bob    | Tuesday | Spinach |      5 |    5.00 |
| Bob    | Tuesday | Cabbage |      5 |    5.00 |
| Cindy  | Monday  | Apple   |      1 |    1.00 |
| Cindy  | Monday  | Orange  |      3 |    3.00 |
| Cindy  | Monday  | Peach   |      3 |    3.00 |
| Cindy  | Monday  | Lettuce |      6 |    6.00 |
| Cindy  | Monday  | Spinach |      8 |    8.00 |
| Cindy  | Monday  | Cabbage |      2 |    2.00 |
| Cindy  | Tuesday | Apple   |      1 |    1.00 |
| Cindy  | Tuesday | Orange  |      3 |    3.00 |
| Cindy  | Tuesday | Peach   |      0 |    0.00 |
| Cindy  | Tuesday | Lettuce |      2 |    2.00 |
| Cindy  | Tuesday | Spinach |      3 |    3.00 |
| Cindy  | Tuesday | Cabbage |      4 |    4.00 |
+--------+---------+---------+--------+---------+

我想将Item 中的值组合成类似的类别，并将信息压缩成更少的行，因为不需要这么多细节。例如，我想将 Apple、Orange 和 Peach 转化为“水果”，将生菜、菠菜和卷心菜转化为“蔬菜”，所有这些都不会丢失存储在 Amount 中的值和Revenue。

我想要实现的表格应该是这样的：

+--------+---------+-----------+--------+---------+
| Vendor |   Day   |   Item    | Amount | Revenue |
+--------+---------+-----------+--------+---------+
| Bob    | Monday  | Fruit     |      4 |    4.00 |
| Bob    | Monday  | Vegetable |     11 |   11.00 |
| Bob    | Tuesday | Fruit     |      4 |    4.00 |
| Bob    | Tuesday | Vegetable |     13 |   13.00 |
| Cindy  | Monday  | Fruit     |      7 |    7.00 |
| Cindy  | Monday  | Vegetable |     16 |   16.00 |
| Cindy  | Tuesday | Fruit     |      4 |    4.00 |
| Cindy  | Tuesday | Vegetable |      9 |    9.00 |
+--------+---------+-----------+--------+---------+

这是我当前的 SQL 查询：

SELECT vendor, 
       day, 
       item_category, 
       Sum(amount), 
       Sum(revenue) 
FROM   (SELECT mytable.*, 
               CASE item 
                 WHEN 'Apple' THEN 'Fruit' 
                 WHEN 'Orange' THEN 'Fruit' 
                 WHEN 'Peach' THEN 'Fruit' 
                 WHEN 'Lettuce' THEN 'Vegetable' 
                 WHEN 'Spinach' THEN 'Vegetable' 
                 WHEN 'Cabbage' THEN 'Vegetable' 
               END AS item_category 
        FROM   mytable) AS x 
GROUP  BY vendor, 
          day, 
          item, 
          item_category;

此查询仍返回完整的行列表，而不是聚合类似的行。我将如何修改它以实现我想要做的事情？

【问题讨论】：

标签： sql postgresql aggregate-functions

【解决方案1】：

您需要从您的 group by 语句中删除项目。下面是更多细节。

由于您正在抓取该项目（苹果、橙子等），因此在您从顶部选择和分组依据中删除该项目之前，它不会仅按 item_category 显示它们。

通过按语句的每个唯一组合进行分组的作品分组，因此必须删除项目。

见下文。注意：我将 item_category 重命名为 item，所以它看起来就像您所说的那样，您希望输出为

SELECT vendor, 
   day, 
   item_category as item, 
   Sum(amount), 
   Sum(revenue) 
FROM   (SELECT mytable.*, 
           CASE item 
             WHEN 'Apple' THEN 'Fruit' 
             WHEN 'Orange' THEN 'Fruit' 
             WHEN 'Peach' THEN 'Fruit' 
             WHEN 'Lettuce' THEN 'Vegetable' 
             WHEN 'Spinach' THEN 'Vegetable' 
             WHEN 'Cabbage' THEN 'Vegetable' 
           END AS item_category 
    FROM   mytable) AS x 
GROUP  BY vendor, 
      day, 
      item_category;

【讨论】：

【解决方案2】：

您必须从嵌套查询中删除项目并按聚合分组：

SELECT vendor, 
   day, 
   item_category, 
   Sum(amount), 
   Sum(revenue) 
FROM   (SELECT vendor, day, amount, revenue,
           CASE item 
             WHEN 'Apple' THEN 'Fruit' 
             WHEN 'Orange' THEN 'Fruit' 
             WHEN 'Peach' THEN 'Fruit' 
             WHEN 'Lettuce' THEN 'Vegetable' 
             WHEN 'Spinach' THEN 'Vegetable' 
             WHEN 'Cabbage' THEN 'Vegetable' 
           END AS item_category 
    FROM   mytable) AS x 
GROUP  BY vendor, 
      day, 
      item_category;

【讨论】：