How to Ignore “Duplicate Key” error in T-SQL (SQL Server)

Question

I have a transaction that contains multiple SQL Statements (INSERT, UPDATE and/or DELETES). When executing, I want to ignore Duplicate Error statements and continue onto the

Answer 1

Use IGNORE_DUP_KEY = OFF during primary key definition to ignore the duplicates while insert. for example

create table X( col1.....)

CONSTRAINT [pk_X] PRIMARY KEY CLUSTERED 
(
)WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON, FILLFACTOR = 70) ON [PRIMARY]
) ON [PRIMARY]

Answer 2

I came here because I was trying to do the same thing; I knew I had dupes in the source data but only wanted to update the target data and not add the dupes.

I think a MERGE works great here because you can UPDATE or DELETE things that are different and INSERT things that are missing.

I ended up doing this and it worked great. I use SSIS to loop through Excel files and load them into a "RAW" SQL table with dupes and all. Then I run a MERGE to merge the "raw" table with the production table. Then I TRUNCATE the "raw" table and move to the next Excel file.

Answer 3

For SQL server 2000:

     INSERT INTO t1 (ID, NAME)
      SELECT valueid, valuename
      WHERE  NOT EXISTS
               (SELECT 0
                FROM   t1 as t2
                WHERE  t2.ID = valueid AND t2.name = valuename)

Answer 4

Expanding on your comment to SquareCog's reply, you could do:

INSERT INTO X VALUES(Y,Z)    WHERE Y  NOT IN (SELECT Y FROM X)
INSERT INTO X2 VALUES(Y2,Z2) WHERE Y2 NOT IN (SELECT Y FROM X2)
INSERT INTO X3 VALUES(Y3,Z3) WHERE Y3 NOT IN (SELECT Y FROM X3)

Here, I assume that column Y is present in all three tables. Note that performance will be poor if the tables are not indexed on Y.

Oh yeah, Y has a unique constraint on it--so they're indexed, and this should perform optimally.

Answer 5

INSERT INTO KeyedTable(KeyField, Otherfield)
SELECT n.* FROM 
    (SELECT 'PossibleDupeLiteral' AS KeyField, 'OtherfieldValue' AS Otherfield
     UNION ALL
     SELECT 'PossibleDupeLiteral', 'OtherfieldValue2'
    )
LEFT JOIN KeyedTable k
    ON k.KeyField=n.KeyField
WHERE k.KeyField IS NULL

This tells the Server to look for the same data (hopefully the same speedy way it does to check for duplicate keys) and insert nothing if it finds it.

I like the IGNORE_DUP_KEY solution too, but then anyone who relies on errors to catch problems will be mystified when the server silently ignores their dupe-key errors.

The reason I choose this over Philip Kelley's solution is that you can provide several rows of data and only have the missing ones actually get in:

Answer 6

Well you could solve this with a temp table..

DECLARE @RoleToAdds TABLE
([RoleID] int, [PageID] int)

INSERT INTO @RoleToAdds ([RoleID], [PageID])
    VALUES
    (1, 2),
    (1, 3),
    (1, 4),
    (2, 5)

INSERT INTO [dbo].[RolePages] ([RoleID], [PageID])
    SELECT rta.[RoleID], rta.[PageID] FROM @RoleToAdds rta WHERE NOT EXISTS 
        (SELECT * FROM [RolePages] rp WHERE rp.PageID = rta.PageID AND rp.RoleID = rta.RoleID)

This might not work for big amounts of data but for a few rows it should work!