CODEWITHSUNDEEP
sqlsql-servert-sql
Tibomso
Tibomso

Reputation: 373

SQL Duplicates optimization

I have the following query:

Original query:

SELECT
    cd1.cust_number_id, cd1.cust_number_id, cd1.First_Name, cd1.Last_Name
FROM @Customer_Data cd1
    inner join @Customer_Data cd2 on
        cd1.Cd_Id <> cd2.Cd_Id
        and cd2.cust_number_id <> cd1.cust_number_id
        and cd2.First_Name = cd1.First_Name
        and cd2.Last_Name = cd1.Last_Name
    inner join @Customer c1 on c1.Cust_id = cd1.cust_number_id
    inner join @Customer c2 on c2.cust_id = cd2.cust_number_id
WHERE c1.cust_number <> c2.cust_number

enter image description here

I optimized it as follows, but there is an error in my optimization and I can't find it:

Optimized query:

    SELECT cd1.cust_number_id, cd1.cust_number_id, cd1.First_Name,cd1.Last_Name
    FROM (
         SELECT cdResult.cust_number_id, cdResult.First_Name,cdResult.Last_Name, COUNT(*) OVER (PARTITION BY cdResult.First_Name, cdResult.Last_Name) as cnt_name_bday  
         FROM @Customer_Data cdResult
         WHERE cdResult.First_Name IS NOT NULL
               AND cdResult.Last_Name IS NOT NULL) AS cd1
    WHERE cd1.cnt_name_bday > 1;

enter image description here

Test data:

DECLARE @Customer_Data TABLE
(
    Cd_Id INT,
    cust_number_id INT,
    First_Name NVARCHAR(30),
    Last_Name NVARCHAR(30)
)

INSERT @Customer_Data (Cd_Id,cust_number_id,First_Name,Last_Name)
VALUES (1, 22, N'Alex', N'Bor'),
       (2, 22, N'Alex', N'Bor'),
       (3, 23, N'Alex', N'Bor'),
       (4, 24, N'Tom', N'Cruse'),
       (5, 25, N'Tom', N'Cruse')

DECLARE @Customer TABLE
(
    Cust_id INT,
    Cust_number INT
)


INSERT @Customer (Cust_id, Cust_number)
VALUES (22, 022),
       (23, 023),
       (24, 024),
       (25, 025)

The problem is that the original query returns 6 rows (duplicating the row). And optimized returns just duplicates, how to make the optimized query also duplicated the row?

Upvotes: 3

Views: 470

Answers (2)

Yogesh Sharma
Yogesh Sharma

Reputation: 50163

You can use only one exists :

SELECT cd.cud_customer_id
FROM dbo.customer_data AS cd
WHERE EXISTS (SELECT 1 
              FROM dbo.customer_data AS c 
              WHERE c.cud_name = cd.cud_name AND c.cud_birthday = cd.cud_birthday AND c.cust_id <> cd.cud_customer_id
             );

Upvotes: 1

Gordon Linoff
Gordon Linoff

Reputation: 1269693

I would suggest just using window functions:

SELECT CD.cud_customer_id
FROM (SELECT cd.*, COUNT(*) OVER (PARTITION BY cud_name, cud_birthday) as cnt_name_bday          FROM dbo.customer_data cd
     ) cd
WHERE cnt_name_bday > 1;

Your query is finding duplicates for either name or birthday. You want duplicates with both at the same time.

Upvotes: 4

Related Questions