I see it now. The trick/bug is that the Connection String is CASE Sensitive! WTF?
This is what I have in place now for Data Source under the Gateway:
If I create a Dataflow for ODBC, and type in a Connection String of "dsn=CData QuickBooks Sys" then I get this, and it associates with an existing Dataset:
But my standard is to capitalize the "DSN", so I normally go for something like "DSN=CDataQuickBooks Sys" and I am offered this:
Note that the Connection name is spelled EXACTLY the same as one that already exists.
The result is this:
Please note how Power BI has NOT figured out that "DSN=..." is the same as "dsn=..." so it created a new one with "(2)" suffix, and ALSO that is then decided to LOWERCASE the "DSN" to make it "dsn".
Not for nothing, but if Connection Strings are CASE SENSITIVE, THEN POWER BI SHOULD NOT BE CHANGING THE CASE OF THE SPELLING!!!
You can't have it both ways. If the Connection String is CASE SENSITIVE, then don't change the case on me, the user. If it is NOT CASE SENSITIVE, then please accept "DSN=..." as a match to "dsn=..."
I still consider this a BUG, but not under the same title!
This user experience is horrible! When setting up a Dataflow, I should be allowed to PICK any existing ODBC Data Source BY NAME. I should NOT be required to spell out the Connection String WITH CASE MATCHING in order for it to figure out it is the same object! And oh, by the way, Power BI RENAMED your connection string behind your back from when you first typed it in and didn't tell you, so if you try to spell out that same connection string exactly as you did before during a subsequent setup, you are going to get a duplicate!
