in the GAE directory emap-dev/waveform-export we still get two seemingly incorrect csns. A search of star_b suggests these patients were never in the source bed, so I don't know why they would show up.
For now I have added the location string to the csv files so that should help with debugging.