Question

I’ve got a set of staging tables where I accept data, scrub, and cleanse them before inserting to the target table. The target table has a primary key constraint and what I’m inserting is a primary key.

I check for the absence of the primary key in the target table before I insert. I insert only records that ARE NOT in the target table, based on the primary key’s absence:

INSERT INTO Target
    SELECT 
        primKey
        , user_state
        , test_state
    FROM
        myStagingTable3

By this point, the stagingTable3 has only data that are NOT present in the Target Table using the following where clause:

WHERE
        primKey not in (Select primKey from Target)

Somehow, I’m getting a primary key violation error :

Msg 2627, Level 14, State 1. (procedure failed at line #…)
Violation of Primary Key constraint ‘pk1101AE.’ Cannot insert
duplicate key in object ‘Target’

My questions:

• under what conditions could a primary key violation occur when the
  key that I want to insert is NOT present in the target table?
• could a prior delete of records cause the primary key to be retained? If so,
  can I work around this somehow?
• something else?

It’s pretty clear that my staging tables have the key and that the Target table does not. Yet the insert fails.

Answer 1

This happens because your staging table contains multiple rows with the same value of primKey.

If it does not matter to you which {user_state, test_state} pair among the duplicated ones makes it into the insert, you can bypass the issue entirely by adding a simple group by, like this:

INSERT INTO Target
    (SELECT 
        primKey
    ,   MAX(user_state)
    ,   MAX(test_state)
    FROM
        myStagingTable3
    GROUP BY primKey)