Deleting rows where the Primary key is duplicated - SQL - sql

Question

My issue is how do we delete a primary key row in case it is duplicated. The other fields may/may not be duplicates. I am interested only in the primary key being duplicated and would like to retain the first instance while deleting the other duplicate entries.
For example,
I have 2 tables with the following data:
Table1:- Portfolio
Columns:- PortfolioID(PK), PortfolioName
Sample data :-
1, North America
2, Europe
3, Asia
Table2:- Account
Columns:- AccountID(PK), PortfolioID(FK), AccountName
Sample data :-
1,1,Quake
1,1,Wind
2,1,Fire
3,1,Quake
4,2,Flood
5,2,Wind
Lets say for PortfolioID = 1,
I am trying to delete row number 2 from the Account table where the AccountID 1 is repeated for PortfolioID =1. I have tried using the CTE expression where I use the ROW_NUMBER statement and try to delete ROWNUMBER <> 1. But this query doesn't work as it deletes all the rows in the table.
The query I tried:
WITH CTE AS
(
SELECT ROW_NUMBER() OVER (PARTITION BY [Account].[AccountID] ORDER BY [Account].[AccountID]) AS [ROWNUMBER],
[Account].[AccountID]
FROM [Account]
INNER JOIN [Portfolio] ON [Portfolio].[PortfolioID] = [Account]. [PortfolioID]
WHERE [Portfolio].[PortfolioID] = 1
)
DELETE [Account]
FROM [CTE]
WHERE [ROWNUMBER] <> 1
Am I doing something wrong in the query? Thanks in advance for the help.

Answer 1

Firstly, if you define the AccountID column as the primary key in your database, this going forward will help solve having these kinds of problems.
Secondly, are you using Sql Server? Which version?
Assuming you are using Sql Server and a recent version which allows you to use windowing, you can try something like this to delete any duplicates that you have.
This will delete ALL copies of ALL duplicates:
WITH CTE AS
(SELECT *,R=RANK() OVER (ORDER BY AccountID,PortfolioID)
FROM Account)
DELETE CTE
WHERE R IN (SELECT R FROM CTE GROUP BY R HAVING COUNT(*)>1)
This alternative script will keep one of the duplicates if that is what you prefer:
WITH CTE AS
(
SELECT *,ROW_NUMBER() OVER (PARTITION BY AccountID,PortfolioID ORDER BY AccountID,PortfolioID) AS RN
FROM Account
)
DELETE FROM CTE WHERE RN<>1
Finally, if you want to only delete duplicates for Portfolio Id 1:
WITH CTE AS
(
SELECT *,ROW_NUMBER() OVER (PARTITION BY AccountID,PortfolioID ORDER BY AccountID,PortfolioID) AS RN
FROM Account
Where PortfolioID = 1
)
DELETE FROM CTE WHERE RN<>1

Answer 2

Primary key column never ever support duplicate entries.
Try with the below query for the desired result based on the given data/inputs.
;WITH CTE AS
(
SELECT ROW_NUMBER() OVER (PARTITION BY a.[AccountID],a.PortfolioID ORDER BY a.[AccountID]) AS [ROWNUMBER],*
FROM [Account] a
WHERE a.[PortfolioID] = 1
)
DELETE
FROM [CTE]
WHERE [ROWNUMBER] > 1