Question

In Oracle, I used to use sequences to generate value for a table’s unique identifier. In a stored procedure, I’d call sequencename.nextval and assign that value to a variable. After that, I’d use that variable for the procedure’s insert statement and the procedure’s out param so I could deliver the newly-generated ID to the .NET client.

I’d like to do the same thing with Teradata, but I am thinking the only way to accomplish this is to create a table that holds a value that is sequentially incremented. Ideally, however, I’d really like to be able to acquire the value that will be used for an identity column’s next value without actually creating a new record in the database.

Answer 1

No, it is not possible with Teradata because Identify values are cached at either the parsing engine (PE) or AMP level based on the type of operation being performed. My understanding is that the DBC.IdCol table shows the next value that will be use to seed the next batch of IDENTITY values that are needed by the PE or AMP.

Another solution would be to avoid using IDENTITY in this manner for your UPI. You could always use the ROW_NUMBER() window aggregate function partitioned by your logical primary key to seed the next range of values for your surrogate key.