如何根据这种特定场景从主表中提取数据？

Question

I am using SQL Server 2014 . I have a table (T1) which contains the list of IDs of cancelled bookings and their equivalent re-bookings.

Extract of table T1 :

CancelledID       Re-bookingID
  301                754
  387                801
  400                900
  ...

Each CancelledID has a unique equivalent in the Re-bookingID column.

I have another table ( T2 ) which contains the list of ALL BookingIDs with additional information related to each ID. An extract of this table is shown below:

BookingID     MonthOfStay      RoomNights
...
301             2019-03-01        10
387             2019-04-01         7
400             2019-03-01         5
754             2019-08-01        10
801             2019-09-01         3
900             2019-07-01         5
900             2019-08-01         4
...   

I need a t-sql query which will me the following output:

  BookingID       Cancelled_MonthOfStay     Re-booking_MonthOfStay     RoomNights
    301                2019-03-01                                           10
    387                2019-04-01                                            7
    400                2019-03-01                                            5
    754                                           2019-08-01                10
    801                                           2019-09-01                 3
    900                                           2019-07-01                 5
    900                                           2019-08-01                 4

As you can see, a re-booking can span over 2 months with additional room nights.

I am thinking about "Joins" between the 2 tables but I am stuck at the logic to be used for the "Joins" (if that is the right way of tackling the problem).

Note: The end in mind is to use this data to build a matrix with Cancelled_MonthOfStay as rows and Re-booking_MonthOfStay as columns. The value inside this matrix will be the total RoomNights.

Any help would be appreciated.

Answer 1

Seems like the easiest method would be to unpivot your data in the " Cancellations " table, and then use a CASE expression to display the data in the appropriate column:

USE Sandbox;
GO

CREATE TABLE dbo.Cancellations (CancelledID int, RebookingID int)
INSERT INTO dbo.Cancellations
VALUES (301,754),
       (387,801),
       (400,900);

GO

CREATE TABLE dbo.Bookings (BookingID int, MonthOfStay date, RoomNights int)
INSERT INTO dbo.Bookings
VALUES (301,'20190301',10),
       (387,'20190401', 7),
       (400,'20190301', 5),
       (754,'20190801',10),
       (801,'20190901', 3),
       (900,'20190701', 5),
       (900,'20190801', 4);
GO

SELECT B.BookingID,
       CASE V.BookingType WHEN 'Cancellation' THEN B.MonthOfStay END AS CancelledMonthOfStay,
       CASE V.BookingType WHEN 'Rebooking' THEN B.MonthOfStay END AS RebookedMonthOfStay,
       B.RoomNights
FROM dbo.Cancellations C
     CROSS APPLY (VALUES(C.CancelledID, 'Cancellation'),(C.RebookingID, 'Rebooking')) V(BookingID,BookingType)
     JOIN dbo.Bookings B ON V.BookingID = B.BookingID
ORDER BY B.BookingID;

GO

DROP TABLE dbo.Cancellations;
DROP TABLE dbo.Bookings;

Answer 2

You can LEFT JOIN the T1 table twice to your T2 table. And then use CASE to leave the right columns empty.

SELECT T2.bookingID,
    case when canc.CancelledID is not null then t2.monthOfSTay end as cancelled_MonthOfStay,
    case when reb.rebookingID is not null then t2.monthOfStay end as rebooking_MonthOfStay

FROM T2
LEFT JOIN T1 as canc
    on T2.bookingID = canc.CancelledID
LEFT JOIN T1 as reb
    on T2.bookingID = reb.rebookingID