Muskrat: repull of data after data was pulled prior to sync leads to duplicate entries for quality score

Spark Interracial Singles had data pulled for Google when it was not in sync,
leading to appearances of Keyword 'blah' in reports as attached report 1 shows
(by design that is). To fix this I ran whsumm which did not help. I ran a sync,
then repulled the data, figuring I needed a bigger hammer. That worked to
associate the data for cost, clicks, and impressions with the correct keyword,
but quality score was not cleared. Ffej tells me this is a bug in the way
quality score is handled in this situation. The problem that led to this is not
that rare right now, but will be once Google sync goes out I hope.

Report 2 shows the final result.