按年份分组的mysql组与年份和月份不同

Question

我正在尝试对我们的订单和销售的产品，收入和总重量做一些报告，但是当我将订单组合在一起时，我的一些查询会得到不同的结果，应该返回相同的结果。

以下是我的疑问：

每月

SELECT
    MONTH(orders.date_purchased) as date,
    YEAR(orders.date_purchased) as year,
    count(DISTINCT orders.orders_id) AS total_orders,
    categories.fields_23 as currency
FROM
orders_shipping_products
Inner Join orders ON orders.orders_id = orders_shipping_products.orders_id
Inner Join categories ON categories.fields_6 = orders.shopping_store_category_id
WHERE
orders.orders_status NOT IN (0, 1, 99)
GROUP BY
date,
year,
categories.fields_23
ORDER by
    YEAR(orders.date_purchased),
    MONTH(orders.date_purchased) ASC,
    categories.fields_23

这将返回2012年的下表，共353个：

+------+------+--------------+----------+
| date | year | total_orders | currency |
+------+------+--------------+----------+
|   11 | 2012 |           86 | EUR      |
|   12 | 2012 |          267 | EUR      |
+------+------+--------------+----------+

每年

SELECT
    YEAR(orders.date_purchased) as year,
    count(DISTINCT orders.orders_id) AS total_orders,
    categories.fields_23 as currency
FROM
orders_shipping_products
Inner Join orders ON orders.orders_id = orders_shipping_products.orders_id
Inner Join categories ON categories.fields_6 = orders.shopping_store_category_id
WHERE
orders.orders_status NOT IN (0, 1, 99)
GROUP BY
year,
categories.fields_23
ORDER by
    YEAR(orders.date_purchased),
    categories.fields_23

它返回2012年的以下

+------+--------------+----------+
| year | total_orders | currency |
+------+--------------+----------+
| 2012 |          351 | EUR      |
+------+--------------+----------+

唯一改变的是total_orders，产品总量，收入和重量是相同的。我每个月检查时只收到两个订单。我还检查了QUARTER，YEAR的选择和分组，返回以下内容：

+---------+------+--------------+----------+
| quarter | year | total_orders | currency |
+---------+------+--------------+----------+
|       4 | 2012 |          351 | EUR      |
+---------+------+--------------+----------+

这让我觉得，当我想每个月生成一份报告时，我可能会对我的选择做错了

WITH ROLLUP

丹尼尔让我尝试在我的查询中添加WITH ROLLUP，按照我所做的并获得回报

SELECT
    MONTH(orders.date_purchased) as date,
    YEAR(orders.date_purchased) as year,
    count(DISTINCT orders.orders_id) AS total_orders,
    categories.fields_23 as currency
FROM
orders_shipping_products
Inner Join orders ON orders.orders_id = orders_shipping_products.orders_id
Inner Join categories ON categories.fields_6 = orders.shopping_store_category_id
WHERE
orders.orders_status NOT IN (0, 1, 99)
GROUP BY
year,
date,
categories.fields_23 WITH ROLLUP

这是返回的数据：

+------+------+--------------+----------+
| date | year | total_orders | currency |
+------+------+--------------+----------+
| 11   | 2012 |           86 | EUR      |
| 11   | 2012 |           86 | NULL     |
| 12   | 2012 |          267 | EUR      |
| 12   | 2012 |          267 | NULL     |
| NULL | 2012 |          351 | NULL     |
+------+------+--------------+----------+

Answer 1

我重新检查了orders_id并且它不是唯一的，实际上有2-3个相同的订单ID，其中一些位于不同月份。发现order_id和另一列的组合构成了真正的唯一ID。将此添加到我的联接中，解决了所有问题。