Compare two number SQL - sql

Question

In SQL,I am trying to compare two numbers in the same field. Both numbers contain different information, but for some technical reason they are same. The problem is when exist one sub-string of length 5 and another of length 4 and the last 4 digits of both are same.I want to get the first one with length 5.
Example:
--------------------------------
|ID | Number| Description |
---------------------------------
| 1 | 12345 | Project X,Ready |
---------------------------------
| 2 | 2345 | Project X,onDesign |
---------------------------------
I should always get 12345(or biggest one) if exist numbers with last 4 digits same. Is there any CASE or CTE statement which can give me an easy resolution for this issue?

Answer 1

Try this:
SELECT Id
,Number
,Description
FROM (
SELECT Id
,Number
,Description
,rank() OVER (PARTITION BY right(cast([Number] AS VARCHAR(20)), 4) ORDER BY Number DESC) AS Ranking
FROM YourTable
) InnerTable
WHERE ranking = 1

Answer 2

Here is an example with not exists:
DECLARE #t TABLE
(
ID INT ,
Number INT ,
Description VARCHAR(100)
)
INSERT INTO #t
VALUES ( 1, 12345, 'Project 1' ),
( 2, 2345, 'Project 2' ),
( 3, 77777, 'Project 3' ),
( 4, 7777, 'Project 4' ),
( 5, 88888, 'Project 5' ),
( 6, 9999, 'Project 6' )
SELECT * FROM #t t1
WHERE NOT EXISTS(SELECT * FROM #t t2
WHERE t2.ID <> t1.ID AND
CAST(t2.Number AS VARCHAR(10)) LIKE '%' + CAST(t1.Number AS VARCHAR(10)))
Output:
ID Number Description
1 12345 Project 1
3 77777 Project 3
5 88888 Project 5
6 9999 Project 6

Answer 3

So you need to join using last 4 digits. You could do this by using simple MOD operator. It's used as a percentage sign in SQL Server.
SELECT 12345 % 10000;
This outputs 2345. Exactly what we are looking for.
So we could build the following query to use that calculation:
DECLARE #Test TABLE
(
ID INT
, Number INT
, Description VARCHAR(500)
);
INSERT INTO #Test(ID, Number, Description)
VALUES (1, 12345, 'Project X,Ready')
, (2, 2345, 'Project X,onDesign');
SELECT T1.*
FROM #Test AS T1
INNER JOIN #Test AS T2
ON T2.Number = T1.Number % 10000
WHERE T2.Number <> T1.Number;
Output:
╔════╦════════╦═════════════════╗
║ ID ║ Number ║ Description ║
╠════╬════════╬═════════════════╣
║ 1 ║ 12345 ║ Project X,Ready ║
╚════╩════════╩═════════════════╝
Note that I've added WHERE T2.Number <> T1.Number. It eliminates equal numbers, because SELECT 2345 % 10000 is 2345 as well.
Update
This could be done using ROW_NUMBER()
;WITH Data (ID, Number, Description, RN)
AS (
SELECT ID
, Number
, Description
, ROW_NUMBER() OVER (PARTITION BY Number % 10000 ORDER BY Number DESC)
FROM #Test
)
SELECT *
FROM Data
WHERE RN = 1;
This will do the classic row_number stuff. It will partition windows by Number % 10000, which means that 12345 and 2345 will fall under same window and the highest number will always come first.

Answer 4

Try this:
SELECT DISTINCT A.*
FROM [Tablename] AS A
INNER JOIN [Tablename] AS B
ON B.Number =RIGHT(A.Number,4)
WHERE B.Number <> A.Number;
RIGHT(A.Number,4) will compare the last 4 digits and will give the output

Answer 5

The query might be RDBMS spesific. For example with MSSQL you can do like this:
SELECT *
FROM myTable AS d1
WHERE NOT EXISTS ( SELECT *
FROM myTable AS d2
WHERE SUBSTRING(d2.number, 2, 4) = d1.number );
EDIT: Ah, you edited and it is an INT! Then you can use the % operator instead of substring.
Sample with CTE:
DECLARE #dummy TABLE
(
id INT IDENTITY
PRIMARY KEY ,
number INT ,
[description] VARCHAR(20)
);
INSERT #dummy ( [number], [description] )
VALUES ( 12345, 'P' ),
( 22345, 'P' ),
( 2345, 'P' ),
( 3456, 'P' ),
( 13456, 'P' ),
( 4567, 'P' );
WITH d AS (
SELECT MAX(number) AS maxNum
FROM #dummy AS [d]
GROUP BY [d].[number] % 10000
)
SELECT d1.*
FROM #dummy AS [d1]
INNER JOIN d ON d.[maxNum] = d1.[number];