SQL查询使用两个表（SQL Server）重叠总和

Question

I'd so appreciate if someone can help with my query.

I have two tables - Payments and Repayments .

Payments:                              Repayments:    
DATE         SUM      IS_REPAYED       DATE         SUM
01/01/15     20 000    0               05/01/15     5 000
                                       06/01/15     10 000

I need to have a flag that shows if Payment sum is covered by Repayment sums, in this case, if (5 000 + 10 000) covers 20 000 , now it's not.

But when new Repayment added, it is:

Payments:                              Repayments:    
DATE         SUM      IS_REPAYED       DATE         SUM
01/01/15     20 000    1               05/01/15     5 000
                                       06/01/15     10 000
                                       07/01/15     5 000

However, there can be a case, when Repayments exceed payment, it means that Payment is repayed and the rest 5 000 (25 000 - 20 000) should be calculated as a Repayment for the next Payment , ie:

Payments:                              Repayments:    
DATE         SUM      IS_REPAYED       DATE         SUM
01/01/15     20 000    1               05/01/15     25 000
01/02/15     10 000    0

and when new repayment comes:

Payments:                              Repayments:    
DATE         SUM      IS_REPAYED       DATE         SUM
01/01/15     20 000    1               05/01/15     25 000
01/02/15     10 000    1               05/02/15     5 000  (or more, ex. 6 000) then 1 000 goes to the next Payment

There is also a case when Repayment comes first, then next payment should have a flag:

Payments:                              Repayments:    
DATE         SUM      IS_REPAYED       DATE         SUM
                                       01/03/15     30 000
05/03/15     30 000     1              

So how can I get this IS_REPAYED flag, when I query the tables at a certain time, using current situation of Payment/Repayment statement? I know I can make a logic for the flag in my application, would it be a better solution?

To have a view I've shown above, I use FULL OUTER JOIN :

SELECT t1.DATE, t1.SUM, NULL AS IS_REPAYED /*?*/, t2.DATE, t2.SUM
FROM 
(   (SELECT t1.*, row_number() OVER (ORDER BY DATE) as seqnum
    FROM Payments t1
    ) t1 
   FULL OUTER JOIN
   (SELECT t2.*, row_number() OVER (ORDER BY DATE) as seqnum
   FROM Repayments t2
   ) t2
ON t1.seqnum = t2.seqnum
)

Answer 1

Hmm I think that you can achieve that with the following sql query

SELECT t1.DATE, SUM(t1.SUM), SUM(t1.SUM) = SUM(t2.SUM) AS IS_REPAYED, t2.DATE, SUM(t2.SUM)
FROM Payments t1 FULL OUTER JOIN Repayments t2 ON t1.DATE = t2.DATE
GROUP BY t1.DATE

Answer 2

1. Updating IS_REPAYED to 0 or 1 in PAYMENTS table when repaymant amount is less or equal to than payment amount for that month.

UPDATE PAYMENTS SET IS_REPAYED = S.IS_REPAYED
FROM
(
    SELECT P.[DATE],CASE WHEN  R.[SUM] = P.[SUM] THEN 1 ELSE 0 END IS_REPAYED
    FROM PAYMENTS P 
    JOIN 
    (
        SELECT CAST(DATEADD(month, DATEDIFF(month, 0, [DATE]), 0)AS DATE) [DATE],  SUM([SUM]) [SUM]
        FROM REPAYMENTS
        GROUP BY CAST(DATEADD(month, DATEDIFF(month, 0, [DATE]), 0)AS DATE)
    )R
    ON DATEPART(MONTH,P.[DATE]) = DATEPART(MONTH,R.[DATE]) AND R.[SUM] <= P.[SUM]
)S
WHERE PAYMENTS.[DATE]=S.[DATE]

• Click here to view result(repaymant amount is less than payment amount)
• Click here to view result(repaymant amount is equal to payment amount)

2. Updating IS_REPAYED to 0 or 1 in PAYMENTS table and insertion to both tables when repaymant amount is greater than payment amount for that month.

Finds sum and groups to each month data in inner query, then compare it with payment table. Since we need this data for further insertions to both PAYMENTS and REPAYMENTS table, we use a temporary table.

SELECT P.[DATE] PAYDATE ,P.[SUM] PAYSUM ,R.[DATE] REDATE,R.[SUM] RESUM
INTO #NEWTABLE
FROM PAYMENTS P 
JOIN 
(
    SELECT  CAST(DATEADD(month, DATEDIFF(month, 0, [DATE]), 0)AS DATE) [DATE],  SUM([SUM]) [SUM]
    FROM REPAYMENTS
    GROUP BY  CAST(DATEADD(month, DATEDIFF(month, 0, [DATE]), 0)AS DATE)
)R
ON DATEPART(MONTH,P.[DATE]) = DATEPART(MONTH,R.[DATE]) AND R.[SUM] > P.[SUM]

Now we find the difference and insert to REPAYMENTS table

INSERT INTO REPAYMENTS
SELECT DATEADD(DAY,1,PAYDATE) [DATE], (RESUM-PAYSUM) REAMOUNT
FROM #NEWTABLE

Similarly we find the difference and insert to PAYMENTS table

INSERT INTO PAYMENTS    
SELECT DATEADD(MONTH,1,PAYDATE) [DATE], (RESUM+(RESUM-PAYSUM))-PAYSUM REAMOUNT,1
FROM #NEWTABLE

Since value in REPAYMENTS table is greater than PAYMENTS table we directly update value of table to 1 from temporary table.

UPDATE PAYMENTS SET ISREPAYED = 1
FROM
(
    SELECT PAYDATE FROM  #NEWTABLE
)N
WHERE PAYMENTS.[DATE]=N.PAYDATE

• Click here to view the result