Why is featurereader reading way more records than are in the input?

Yeah, this is a common mistake...

As the documentation mentions, a complete read is done for each feature that enters the Initiator port. 

The Initiator exit port shows there are 218 initiating objects processed (so far), so every record in the input data set has been read 218 times (so far).

You should have only one feature enter the FeatureReader.

Thank you. Can you select for a single record in the input? ie 'uniqueid = 1'?

You can use a Sampler, sampling only the first feature that enters.

Or you can place the FeatureReader after a Creator.

Or maybe (depending on the input source) you can replace the Creator and the FeatureReader with a SQLCreator.

Thank you. I appreciate it. Yes, it was reading each record 1681 times. A where clause that selected a single records seemed to work.

I've tried scenario one and the 'generic' output, which has the correct number of records, has no fields '<no schema>'

Sampler and specifying the output port seems to do the trick. I really appreciate your help.

Ah yes, I created a quick mockup for the visualization. You should indeed specify the output port.

I'm glad I could help you.