BULK INSERT ignores UNIQUE index with IGNORE_DUP_KEY set?

All,

Just want to make sure that I understand what's going on here.

I have a table with IGNORE_DUP_KEY set on a unique, multi-column
index.

What I'm seeing is this:

1) When performing a BULK INSERT, the UNIQUE index is not being
respected and rows which violate the unique index are inserted.

2) When performing a regular INSERT, the UNIQUE index is being
respected and rows which violate the unique index ARE NOT inserted.

Is this expected behavior.

Also, I have some questions, given the index described.

Q1) Will a regular INSERT that attempts to insert duplicate data get
an error back or just a warning?

Q2) How can I set things up so that a BULK INSERT would NOT allow
duplicates to be entered into the table?

Thanks,
Wes Gamble

Re: BULK INSERT ignores UNIQUE index with IGNORE_DUP_KEY set?

On Jun 22, 5:33 pm, Weyus wrote:
> All,
>
> Just want to make sure that I understand what's going on here.
>
> I have a table with IGNORE_DUP_KEY set on a unique, multi-column
> index.
>
> What I'm seeing is this:
>
> 1) When performing a BULK INSERT, the UNIQUE index is not being
> respected and rows which violate the unique index are inserted.
>
> 2) When performing a regular INSERT, the UNIQUE index is being
> respected and rows which violate the unique index ARE NOT inserted.
>
> Is this expected behavior.
>
> Also, I have some questions, given the index described.
>
> Q1) Will a regular INSERT that attempts to insert duplicate data get
> an error back or just a warning?
>
> Q2) How can I set things up so that a BULK INSERT would NOT allow
> duplicates to be entered into the table?
>
> Thanks,
> Wes Gamble

I made a logical error that I just realized. The BULK INSERT that I
was doing, in fact did have unique enough data.

However, my question Q1 still stands - what is the behavior of an
INSERT of duplicate data against this type of index - I suspect that
you get a warning
and not an error - is that correct?

Thanks,
Wes

Re: BULK INSERT ignores UNIQUE index with IGNORE_DUP_KEY set?

Weyus (weyus@att.net) writes:
>> Q1) Will a regular INSERT that attempts to insert duplicate data get
>> an error back or just a warning?
>>...
> However, my question Q1 still stands - what is the behavior of an
> INSERT of duplicate data against this type of index - I suspect that
> you get a warning and not an error - is that correct?

Yes, SQL Server only emits a warning. However, there is a catch, some
client APIs incorrectly interprets this as an error. If run:

CREATE TABLE blamblam (a int NOT NULL)
CREATE UNIQUE INDEX updix ON blamblam(a) WITH IGNORE_DUP_KEY
go
EXEC master..xp_cmdshell 'ECHO 12 > C:\temp\blamblam.txt'
EXEC master..xp_cmdshell 'ECHO 12 >> C:\temp\blamblam.txt'
go
BULK INSERT blamblam FROM 'C:\temp\blamblam.txt'
go
SELECT * FROM blamblam
go
DROP TABLE blamblam

I get this output from Mgmt Studio:

Duplicate key was ignored.

(1 row(s) affected)
a
-----------
12

(1 row(s) affected)

But from Query Analyzer, against the same server instance, I get:


Server: Msg 3604, Level 16, State 1, Line 1
Duplicate key was ignored.

a
-----------
12
(1 row(s) affected)


--
Erland Sommarskog, SQL Server MVP, esquel@sommarskog.se

Books Online for SQL Server 2005 at
http://www.microsoft.com/technet/prodtechnol/sql/2005/downlo ads/books.mspx
Books Online for SQL Server 2000 at
http://www.microsoft.com/sql/prodinfo/previousversions/books .mspx