Thanks to those who responded.

It's looking as though this will be simpler for me than expected, in that I will need only historical rates – and these I can load...

That looks fine to me.

How are you consuming/viewing the results of the query in SSIS?

Frederico is spot on.

In addition, note that SSDT is a client tool and should, therefore, not be installed on either of your servers.

• This reply was modified 5 years, 12 months ago by Phil Parkin. Reason: Fix typo

When you created the Project in VS, what sort of Project did you create? Was it like this:

limitations of using developer edition versus full blow standard?

As John suggested, quite the opposite. Developer has more features than Standard.

No need for a cte:

update h
   set h.Person = t.Person_To, 
       h.Pctg = t.Pctg_New 
  from #holders...

Please bear with me.  I've never had to create a project in the 23 years I've worked with SSIS.  I created a SQL Server project and add my .dtsx...

Hi Phil, thanks for replying. This is exactly what I thought.  This is all at a POC level so far.

I originally had my dimensional loads running in sequence but...

Yuck! I started on this before spotting the need for some sort of recursion, to make the Person A4 bit work.

If I had more time, I think that I would...

I'm with you on error message clarity. On Azure,...

You need to use a data type of Object to handle resultsets.

Sounds like you need to put your data flow task inside a Foreach container.

Your Foreach container needs to iterate around the various GLExtractId values.

To make that happen, use an ExecuteSQL...

You have two (or more) child packages running in parallel.

Your requirement is that if one of the child packages fails, processing stops immediately ... is that correct?

If so, I don't...

Looks good to me ... that should work (assuming you have no values for NodeId which are less than -9999 ... which you should confirm).