That looks like the sort of error you would get if you copied a package containing a project connection from one project to another.

You are right to consider the potential for duplicates. I can't think of an elegant (and fool-proof) way of handling the composite key – so I'd stick with Jonathan's suggestion.

If your actual tables include a PK not mentioned here, you may be able to use that to make your row unique, which would obviate the need for a derived...

It's a Kerberos thing. Get yourself a big cup of tea and read...

You can...

It's a bit hacky, but this avoids the UNION ALL:

WITH Uniqued
AS (SELECT Grouper = IIF(sd.SampleType = 1, 1, CHECKSUM (*))
       ...

Try

SUM(IIF(SampleType = 1, SampleValue,0))

Like this?

WITH qry1
AS (SELECT *
    FROM
    (
        VALUES
      ...

How many tables are updated during this process?

Does each table have date created / date modified columns?

Is the incoming data purely inserts, or inserts + updates, or inserts + updates...

Your final resultset cannot contain a mixture of numbers of columns and datatypes. But you can create a wider dataset, eg

SELECT Col1, Col2, Col3, Col4 = NULL,...

Jeff's post says exactly what he meant to say. He did not condense his...

A quick response to some of your questions:

• All environment-specific items should be parameterised.
• Don't Save Sensitive, along with sensitive parameters whose values are passed from SSISDB environments, is the...

OK. What is your question?

That's...

Check out COUNT(x) vs COUNT(DISTINCT x)