我在使用SQL查询时遇到了一些困难。表名为Customer_Orders，包含以下列
Order_ID INT -Primary Key
Customer_ID INT - Customer who made the transaction
Order_Time_Stamp DATE - Date and time of transaction
Item STRING - Description of item purchased
Transaction_Status STRING - Status can be "Success" and "Failure"
如果第一个订单成功的客户，我们怎样才能在30天内获得每个客户的第一个订单和第二个订单之间的持续时间？
我的查询逻辑类似于：
select Customer_ID, Duration_Between_Order_In_Days
from Customer_Orders
where DATEDIFF(DAYS, First_Order_Date, Second_Order_Date) < 30
and First_Order_Status='Success'
我尝试过使用DATEDIFF（）函数和自联接和子查询，但我不知道如何为每个客户找到第二个订单。任何想法将不胜感激！
答案 0 :(得分：0)
select Customer_ID, Duration_Between_Order_In_Days
from Customer_Orders
where DATEDIFF(First_Order_Date, Second_Order_Date) < 30
and First_Order_Status='Success'
如上面的查询所示，试试这个。
只是为了关心您的First_Order_Date, Second_Order_Date必须采用'YYYY-MM-DD'的形式。
希望这会对你有所帮助。
答案 1 :(得分：0)
使用adventureworks2012 [Sales]。[SalesOrderHeader]前10条记录如下（在将客户11000的第一个salesorderid的状态重置为0之后），其中5表示成功
+--------------+---------------------+------------+--------+
| salesorderid | orderdate | customerid | status |
+--------------+---------------------+------------+--------+
| 1 | 2009-10-13 00:00:00 | 0 | 0 |
| 43793 | 2005-07-22 00:00:00 | 11000 | 0 |
| 51522 | 2007-07-22 00:00:00 | 11000 | 5 |
| 57418 | 2007-11-04 00:00:00 | 11000 | 5 |
| 43767 | 2005-07-18 00:00:00 | 11001 | 5 |
| 51493 | 2007-07-20 00:00:00 | 11001 | 5 |
| 72773 | 2008-06-12 00:00:00 | 11001 | 5 |
| 43736 | 2005-07-10 00:00:00 | 11002 | 5 |
| 51238 | 2007-07-04 00:00:00 | 11002 | 5 |
| 53237 | 2007-08-27 00:00:00 | 11002 | 5 |
+--------------+---------------------+------------+--------+
10 rows in set (0.18 sec)
Mysql的
select f.customerid,f.firstorder,o2.`Status`,s.secondorder,o3.`status`,
o2.OrderDate, o3.OrderDate, datediff(o3.OrderDate, o2.OrderDate) daysbetweenorders
from
(
select customerid,min(salesorderid) firstorder
from orders
where #customerid = 11000 and
status = 5
group by customerid
) f
join
(
select customerid,min(o.salesorderid) secondorder
from orders o
where #o.customerid = 11000 and
o.salesorderid > (select min(salesorderid) from orders o1 where o1.customerid = o.customerid and o1.Status = 5)
group by customerid
) s on s.customerid = f.customerid
join orders o2 on o2.SalesOrderID = f.firstorder
join orders o3 on o3.SalesOrderID = s.secondorder
limit 10
结果
+------------+------------+--------+-------------+--------+---------------------+---------------------+-------------------+
| customerid | firstorder | Status | secondorder | status | OrderDate | OrderDate | daysbetweenorders |
+------------+------------+--------+-------------+--------+---------------------+---------------------+-------------------+
| 11000 | 51522 | 5 | 57418 | 5 | 2007-07-22 00:00:00 | 2007-11-04 00:00:00 | 105 |
| 11001 | 43767 | 5 | 51493 | 5 | 2005-07-18 00:00:00 | 2007-07-20 00:00:00 | 732 |
| 11002 | 43736 | 5 | 51238 | 5 | 2005-07-10 00:00:00 | 2007-07-04 00:00:00 | 724 |
| 11003 | 43701 | 5 | 51315 | 5 | 2005-07-01 00:00:00 | 2007-07-09 00:00:00 | 738 |
| 11004 | 43810 | 5 | 51595 | 5 | 2005-07-26 00:00:00 | 2007-07-26 00:00:00 | 730 |
| 11005 | 43704 | 5 | 51612 | 5 | 2005-07-02 00:00:00 | 2007-07-27 00:00:00 | 755 |
| 11006 | 43819 | 5 | 51198 | 5 | 2005-07-27 00:00:00 | 2007-07-02 00:00:00 | 705 |
| 11007 | 43743 | 5 | 51581 | 5 | 2005-07-12 00:00:00 | 2007-07-25 00:00:00 | 743 |
| 11008 | 43826 | 5 | 51282 | 5 | 2005-07-28 00:00:00 | 2007-07-07 00:00:00 | 709 |
| 11009 | 43837 | 5 | 51562 | 5 | 2005-07-30 00:00:00 | 2007-07-24 00:00:00 | 724 |
+------------+------------+--------+-------------+--------+---------------------+---------------------+-------------------+
10 rows in set (0.77 sec)
Sql server 非常相似，但状态需要转义，datediff语法不同，top替换限制
select top 10 f.customerid,f.firstorder,o2.[Status],s.secondorder,o3.[status],
o2.OrderDate, o3.OrderDate, datediff(d,o2.OrderDate, o3.OrderDate) daysbetweenorders
from
(
select customerid,min(salesorderid) firstorder
from [Sales].[SalesOrderHeader]
where --customerid = 11000 and
status = 5
group by customerid
) f
join
(
select customerid,min(o.salesorderid) secondorder
from [Sales].[SalesOrderHeader] o
where --o.customerid = 11000 and
o.salesorderid > (select min(salesorderid) from [Sales].[SalesOrderHeader] o1 where o1.customerid = o.customerid and o1.Status = 5)
group by customerid
) s on s.customerid = f.customerid
join [Sales].[SalesOrderHeader] o2 on o2.SalesOrderID = f.firstorder
join [Sales].[SalesOrderHeader] o3 on o3.SalesOrderID = s.secondorder
order by f.customerid
注意： - 确保你有合适的密钥 - 我在下面看到了mysql（不能保证这些是你的应用程序）。
CREATE TABLE `orders` (
`SalesOrderID` int(11) NOT NULL,
`RevisionNumber` int(11) NOT NULL,
`OrderDate` varchar(50) NOT NULL,
`DueDate` varchar(50) NOT NULL,
`ShipDate` varchar(50) DEFAULT NULL,
`Status` int(11) NOT NULL,
`OnlineOrderFlag` int(11) NOT NULL,
`SalesOrderNumber` int(11) DEFAULT NULL,
`PurchaseOrderNumber` int(11) DEFAULT NULL,
`AccountNumber` int(11) DEFAULT NULL,
`CustomerID` int(11) NOT NULL,
`SalesPersonID` int(11) DEFAULT NULL,
`TerritoryID` int(11) DEFAULT NULL,
`BillToAddressID` int(11) NOT NULL,
`ShipToAddressID` int(11) NOT NULL,
`ShipMethodID` int(11) NOT NULL,
`CreditCardID` int(11) DEFAULT NULL,
`CreditCardApprovalCode` varchar(15) DEFAULT NULL,
`CurrencyRateID` int(11) DEFAULT NULL,
`SubTotal` decimal(10,2) NOT NULL,
`TaxAmt` decimal(10,2) NOT NULL,
`Freight` decimal(10,2) NOT NULL,
`TotalDue` decimal(10,2) DEFAULT NULL,
`Comment` varchar(128) DEFAULT NULL,
`rowguid` varchar(50) DEFAULT NULL,
`ModifiedDate` varchar(50) NOT NULL,
KEY `k001` (`OrderDate`,`SalesOrderID`),
KEY `k1` (`CustomerID`,`SalesOrderID`),
KEY `k2` (`SalesOrderID`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1