Find the difference between top 2 rows - sql

Question

I am trying to workout the difference between the top 2 rows for all columns in a table. I will just specify one column to make this easier. I am new to SQL writing so apologies if this is easy. I am on SSMS and so far i think i need to inner join the table then compare row 1 of table1 with row 2 of table2. The id column works in that the newest row is the highest id. I need to take the highest id and then the second highest id (row 2) and find the difference.
SELECT Table1.id,
Table1.transferdate,
Table1.payment,
Table2.id,
Table2.transferdate,
Table2.payment
FROM Table1 AS Table1
INNER JOIN Table2 AS Table2 ON Table1.id = Table2.id
I want to see the difference between yesterday (Top row) and previous day to that (second row) in the payment column. The payment column should ever increase as that data is added on to the previous days. I'm just not sure where to go with it after the INNER JOIN and nothing I have tried has worked.
Data example of what i currently have:
id | transferdate |payment | debt | mailing_batch
46 | 2017-05-18 |651681 | 616816 | 1861651
45 | 2017-05-17 |601680 | 516168 | 1616866
What i want is the difference:
id | transferdate |payment | debt | mailing_batch
1 | 1 |50001 | 100648 | 244785
I only ever want to see the difference between the top 2 for each column. Would i just delete out the select column names and just leave the ones that i have used the LEAD function on with TOP1?
Not interested in any other of the rows. Just always the top 2 as this is a data copy table and its a way to ensure the data has updated correctly in the business set up.

Answer 1

SELECT Top 1 Table1.id,
Table1.payment-LEAD(Table1.payment,1) over(order by Table1.id desc)
FROM Table1 AS Table1
INNER JOIN Table2 AS Table2 ON Table1.id = Table2.id
order by Table1.id desc
If you are using a version of sql server 2012 or higher, you can use the above query.

Answer 2

;With CTE(id , transferdate ,payment , debt , mailing_batch)
AS
(
SELECT 46 , '2017-05-18' ,651681 , 616816 , 1861651 UNION ALL
SELECT 45 , '2017-05-17' ,601680 , 516168 , 1616866
)
SELECT Id
,transferdate
,payment
,debt
,mailing_batch
FROM (
SELECT (LEadId - Id) AS Id
,DATEDIFF(DAY, transferdate, Leadtransferdate) AS transferdate
,(Leadpayment - payment) AS payment
,(Leaddebt - debt) AS debt
,(Leadmailing_batch - mailing_batch) AS mailing_batch
,ROW_NUMBER() OVER (
ORDER BY (
SELECT 1
)
) AS Seq
FROM (
SELECT *
,LEAD(id) OVER (
ORDER BY id
) AS LEadId
,LEAD(transferdate) OVER (
ORDER BY transferdate
) AS Leadtransferdate
,LEAD(payment) OVER (
ORDER BY payment
) AS Leadpayment
,LEAD(debt) OVER (
ORDER BY debt
) AS Leaddebt
,LEAD(mailing_batch) OVER (
ORDER BY mailing_batch
) AS Leadmailing_batch
FROM CTE
) Dt
) Final
WHERE Final.Seq = 1
OutPut
id | transferdate |payment | debt | mailing_batch
1 | 1 |50001 | 100648 | 244785

Answer 3

Which version of sql server you're using ? If it is 2012+ then LEAD and LAG functions will work. You can read more about them here and here
Try it yourself as they're really easy. Show some efforts and if not working, I will help you out.
update after comments
CREATE TABLE [dbo].[t1_new](
[Id] [int] NULL,
[transferdate] [date] NULL,
[payment] [int] NULL
) ON [PRIMARY]
GO
INSERT [dbo].[t1_new] ([Id], [transferdate], [payment]) VALUES (1, CAST(N'2017-05-18' AS Date), 1000)
GO
INSERT [dbo].[t1_new] ([Id], [transferdate], [payment]) VALUES (2, CAST(N'2017-05-19' AS Date), 1100)
GO
INSERT [dbo].[t1_new] ([Id], [transferdate], [payment]) VALUES (3, CAST(N'2017-05-20' AS Date), 1200)
GO
INSERT [dbo].[t1_new] ([Id], [transferdate], [payment]) VALUES (4, CAST(N'2017-05-21' AS Date), 1400)
GO
select top 1 t1.*,
LAG(transferdate) OVER (ORDER BY transferdate) previous_date,
LAG(payment) OVER (ORDER BY payment) previou_payment,
t1.payment - (LAG(payment) OVER (ORDER BY payment)) as payment_difference
from t1_new t1
order by Id desc
screenshots :
input table
output table
PS: you don't need 2 separate tables. Your issue can be resolved by 1 table itself.