Question

当前数据：

Status table
CustomerID    Event      StartDateTime
1001          check-in   2015-04-13 10:15
1001          Transfer   2015-04-13 10:30
1001          check-in   2015-04-13 10:45
1001          check-out  2015-04-13 12:00
2001          Transfer   2015-04-15 12:00
2001          check-out  2015-04-15 12:30
3001          WaitRoom   2015-05-16 10:00
3001          check-in   2015-05-16 10:15

Location table
CustomerID    Location
1001          River
2001          Mountain
3001          Sun Rise

Facilty table
Customer ID     Description
1001            Gym
2001            Pool
3001            Breakfast

Required result:
CustomerID  Event     StartDateTime     Location     Description
1001        check-in  2015-04-13 10:15  River        Gym
2001        Transfer  2015-04-15 12:00  Mountain     Pool
3001        check-in  2015-05-16 10:15  Sun Rise     Breakfast

根据第一次登记或转移事件挑选最早的StartDateTime。从首次办理登机手续获取StartDateTime。如果没有办理登机手续，请先获取Transfer StartDateTime。非常感谢。更新：以下是我正在尝试但它仍然没有产生正确的结果。 http://sqlfiddle.com/#!6/18d45/18

SELECT a.CustomerID, coalesce(b.Event,c.Event), coalesce(b.StartDateTime,c.StartDateTime),d.Location,e.Description
FROM (SELECT DISTINCT CustomerID FROM Status) a
LEFT JOIN Status b ON a.CustomerID = b.CustomerID AND b.Event='check-in'
LEFT JOIN Status c ON a.CustomerID = c.CustomerID AND c.Event='Transfer'
INNER JOIN Location d on a.CustomerID=d.CustomerID 
INNER JOIN Facility e on e.CustomerID=d.CustomerID

Answer 1

WITH stat AS (
SELECT CustomerID
, Event
, StartDateTime
, ROW_NUMBER() OVER (PARTITION BY CustomerID ORDER BY StartDateTime) as RowN
FROM Status
)
SELECT s.CustomerID, s.Event, s.StartDateTime, l.Location, f.Description
FROM stat AS s
JOIN Location AS l ON s.CustomerID = l.CustomerID
JOIN Facilty AS f ON s.CustomerID = f.CustomerID
WHERE s.RowN = 1

希望有所帮助：）

Answer 2

您可以使用ROW_NUMBER：

执行此操作

SQL Fiddle

WITH Cte AS(
    SELECT *,
        RN = ROW_NUMBER() OVER(PARTITION BY CustomerID ORDER BY
                                CASE
                                    WHEN Event = 'check-in' THEN 1
                                    ELSE 2
                                END,
                                StartDateTime
                            )
    FROM Status
    WHERE Event IN('check-in', 'Transfer')
)
SELECT
    c.CustomerID, c.Event, c.StartDateTime, l.Location, f.Description
FROM Cte c
INNER JOIN Location l
    ON l.CustomerID = c.CustomerID
INNER JOIN Facility f
    ON f.CustomerID = c.CustomerID
WHERE RN = 1

不使用ROW_NUMBER：

SQL Fiddle

SELECT
    s.CustomerID,
    CASE WHEN MaxCheckIn IS NOT NULL THEN 'check-in' ELSE 'Transfer' END AS Event,
    CASE WHEN MaxCheckIn IS NOT NULL THEN MaxCheckIn ELSE MaxTransfer END AS StartDateTime,
    l.Location,
    f.Description
FROM (
    SELECT
        CustomerID,
        MAX(CASE WHEN Event = 'check-in' THEN StartDateTime END) AS MaxCheckIn,
        MAX(CASE WHEN Event = 'Transfer' THEN StartDatetime END) AS MaxTransfer
    FROM Status
    WHERE Event IN ('check-in', 'Transfer')
    GROUP BY CustomerID
)s
INNER JOIN Location l
    ON l.CustomerID = s.CustomerID
INNER JOIN Facility f
    ON f.CustomerID = s.CustomerID

Answer 3

这可以通过指定虚拟等级（根据您的逻辑更改它以匹配）然后row_number到您的等级来处理。然后，只需在out_number等于1的外部查询中过滤掉以获得所需的输出：

;WITH stat AS (
SELECT CustomerID
, Event
, StartDateTime
, CASE 
    WHEN event = 'check-in' then 1
    WHEN event = 'Transfer' then 2
  ELSE 3
  END  as Ranking
FROM Status
),
q2 as
(
SELECT 
    s.*,
    ROW_NUMBER () over (partition by customerid order by s.ranking) as rn
FROM stat AS s
)
select q2.customerid,
       q2.[event],
       q2.startdatetime,
       l.location,
       f.[description]
 from q2
JOIN Location AS l ON q2.CustomerID = l.CustomerID
JOIN Facility AS f ON q2.CustomerID = f.CustomerID
WHERE q2.rn = 1

SQL Fiddle Demo

根据不同的事件选择最早的日期并加入其他表

3 个答案: