按日期计算新行和缺失行数

Question

我有这样的表结构。

+--------------------------+------------------------+
|  id |    date      | property_id | - other cols - |
+--------------------------+------------------------+
|  1  |   2012-12-04 |    102      |
|  2  |   2012-12-05 |    101      |
|  3  |   2012-12-05 |    102      |
|  4  |   2012-12-05 |    103      |
|  5  |   2012-11-07 |    101      |
|  6  |   2012-11-07 |    102      |
|  7  |   2012-11-07 |    103      |
|  8  |   2012-10-08 |    101      |
|  9  |   2012-10-08 |    102      |
|  10 |   2012-10-08 |    103      |
|  11 |   2012-10-08 |    104      |
+-----+--------------+-------------+-----------------+

这是一张跟踪房产价格和其他历史数据的历史表。 id是唯一ID，property_id随着时间的推移有多个条目。每周进行历史数据的批量输入。例如，对于两个不同的日期，我们具有相同的属性ID，表示两个日期的价格变化。

现在我需要的是找到特定的给定时间段（日期之间）。

1. 与上一个输入日期相比，新增属性的数量。
2. 上一个输入日期缺少的属性数量。

我希望我的解释足够清楚。

修改

将前一周的＃2更改为上一个输入日期。

Answer 1

对于第一个问题，您可以使用以下查询：

     Select count(distinct(property_id)) from table-name 
      where date Between start_date AND end_date 
      AND 
      property_id != ALL
      (
            Select distinct(property_id)
            from table-name 
            where date < start_date
      )

这将在start_date日期之后显示所有新属性_d，直到end_date

对于第二个，你可以使用它：

   Select count(distinct(property_id)) from table-name 
      where date < start_date
      AND 
      property_id != ALL
      (
            Select distinct(property_id)
            from table-name 
            where date Between start_date AND end_date
      )

这将显示start_date之前的所有property_id，它们在start_date和end_date中不存在

Answer 2

Sql Fiddle

假设，表名为orders

与上一个输入日期相比，添加了新属性的数量。

SELECT 
  SUM(total.property_id_count) as total_property
FROM
  (
    SELECT   
      count(all_order.property_id) as property_id_count,
      date_order.property_id  from orders as all_order
    RIGHT JOIN 
      (SELECT orders.* FROM orders WHERE `date` BETWEEN $start_date  AND $end_date) as   date_order
ON all_order.id = date_order.id
GROUP BY all_order.property_id 
HAVING COUNT(all_order.property_id) > 1) AS total

上周失踪的物业数量。

SELECT 
  SUM(total.property_id_count) as total_property
FROM
  (
    SELECT date_order.date,   
     count(all_order.property_id) as property_id_count,
     date_order.property_id  from orders as all_order
    RIGHT JOIN 
     (
       SELECT orders.* FROM orders
       where date between date_sub(now(),INTERVAL 1 WEEK) and now()
     ) as date_order
ON all_order.id = date_order.id
GROUP BY all_order.property_id 
HAVING COUNT(all_order.property_id) > 1

) as total