How do I use an INSERT statement's OUTPUT clause to get the identity value?

Question

If I have an insert statement such as:

INSERT INTO MyTable
(  
  Name,
  Address,
  PhoneNo
)
VALUES
(
  'Yatrix',
   '1234 Address Stuff',
   '1112223333'
)

How do I set @var INT to the new row's identity value (called Id) using the OUTPUT clause? I've seen samples of putting INSERTED.Name into table variables, for example, but I can't get it into a non-table variable.

I've tried OUPUT INSERTED.Id AS @var, SET @var = INSERTED.Id, but neither have worked.

Jevons

I know about @@SCOPE_IDENTITY already, I specifically want to know how to do it with OUPUT. Thanks. – Jevons

You need to insert it into a table variable then select from that. There is no syntax to assign directly to a scalar variable from the OUTPUT clause. – Feltonfelts

The OUTPUT clause has to output into a table or table variable.. – Algernon

The OUTPUT clause writes to a table. It can be a table variable, temporary table, ... . – Clarisaclarise

marc_s answer is correct but this isn't needed to get the identity of the last inserted. I agree to avoid the @@IDENTITY because of trigger issues. There is a built-in function just for this case called SCOPE_IDENTITY(). Simply replace @@IDENTITY with SCOPE_IDENTITY() to get the truly inserted identity. – Medicate

My question specifically asks for the OUTPUT clause. – Jevons

Answer 1

You can either have the newly inserted ID being output to the SSMS console like this:

INSERT INTO MyTable(Name, Address, PhoneNo)
OUTPUT INSERTED.ID
VALUES ('Yatrix', '1234 Address Stuff', '1112223333')

You can use this also from e.g. C#, when you need to get the ID back to your calling app - just execute the SQL query with .ExecuteScalar() (instead of .ExecuteNonQuery()) to read the resulting ID back.

Or if you need to capture the newly inserted ID inside T-SQL (e.g. for later further processing), you need to create a table variable:

DECLARE @OutputTbl TABLE (ID INT)

INSERT INTO MyTable(Name, Address, PhoneNo)
OUTPUT INSERTED.ID INTO @OutputTbl(ID)
VALUES ('Yatrix', '1234 Address Stuff', '1112223333')

This way, you can put multiple values into @OutputTbl and do further processing on those. You could also use a "regular" temporary table (#temp) or even a "real" persistent table as your "output target" here.

Fenian

The answer here for the code behind was concise. ExecuteScalar() FTW – Arbutus

You can insert the result in a real persistent table - this is extremely fantastic because it means that you can INSERT information in TWO tables at the same time. – Unplumbed

Don't ever use @@IDENTITY to pull from the top. Foud out the hard way working with triggers and since they were recording history of changes made to one table and inserting into a new table at the same time @@IDENTITY started returning back values from the history table. hilarity ensues from there! Please use marc_s' solution. for the time being I have went with the @OutputTbl method, but I'm intrigued by the other options. – Shakiashaking

OUTPUT INTO is extremely fantastic except that "The target table of the OUTPUT INTO clause cannot be on either side of a (primary key, foreign key) relationship", which for me is about 99% of potential use cases. I assume this is because the OUTPUT clause can return data even when the transaction is rolled back, but it's a little annoying it is so hard to insert into data into related tables A and B in one shot. – Mitchellmitchem

@EricBishard SCOPE_IDENTITY() works better for that. – Ithnan

@DerreckDean SCOPE_IDENTITY() seems to return the last identity before the insert, which may be from a different table completely. I've made sure there are no triggers on the INSERT target table, although even with a trigger inserting another record into another table with an identity field, the results are the same. E.g. If I have OUTPUT inserted.ID, SCOPE_IDENTITY() INTO @OutputTable, the two values are different. After the second insert, ID has a value one greater than SCOPE_IDENTITY – Whiles

In the case of foreign key constraints (the norm, ha ha Microsoft), I sometimes add another column to the target table to receive the source identity field, so I can later join it back. This has an auditing advantage in my applications as well. – Whiles

I believe it is scoped to your connection, and in my experience I have never had it return an ID from a different operation. YMMV, though. – Ithnan

@Yogurtu: the OUTPUT .... INTO ... only supports tables - even with just one column - but it cannot be a scalar variable – Fenian

Not tested this, but I'm betting that why you @DaveBoltman are seeing different numbers is that the SCOPE_IDENTIY() isn't set until after the insert has completed. Thus, you are seeing the value from the previous one (as expected - if you did something like insert into foo (some_field) values (SCOPE_IDENTITY()) - you wouldn't expect that to give you the value you just inserted now would you? No, during the query, it is the value from the previous query.) . Never had SCOPE_IDENTITY() work unexpectedly, although, I LOVE the idea of using Output, and it is awesome in merge commands – Encaustic

Never use @@identity. Never use scope_identity(). output into is the only way to correctly, atomically, get the identity of the last inserted record. – Became

@marc_s, the solution with @OutputTbl was the first solution I've tried (following MS's docs). However, the returned ID is 0 always. My table has an INSTEAD OF IUD trigger, which inserts and updates actual values but does not use the idenity column (ID) in those operations as a destination. Any suggestions? – Affusion

@Robert Calhoun use MERGE for that scenario – Antoinette

Answer 2

I didn't like the 'trust me, it's the last ID somewhere' element of the above, so browsed Learn.Microsoft and with a bit of jiggling got a neat and exact solution, informed by Micrsosoft examples. (You might not like the old-man-style code, but hey, whatever). It's a C# function but you can extract the SQL elements as required.

const string queryString =
    "DECLARE @MyTableVar TABLE(NewParaID INT); "
    + "INSERT into dbo.Para(ParaText) "
    + "OUTPUT INSERTED.ParaID INTO @MyTableVar "
    + "VALUES(@para); "
    + "SELECT NewParaID FROM @MyTableVar; ";

using (SqlConnection connection =
new(m_sConnectionString))
{
    SqlCommand command = new(queryString, connection);
    command.Parameters.AddWithValue("@para", sPara);
    
    try
    {
        connection.Open();

        nParaID = (int)command.ExecuteScalar();
    }
    catch (Exception ex)
    {
        sError = ex.Message;
        bResult = false;
    }
}

Store