We were keeping track of the number of statistical summaries per
entry in the "composition" field. When indexing that structure with UUIDs,
we can run into a situation where if there are duplicate UUIDs in the
composition field, the index into the result Timeseries structure becomes off.
We were keeping track of the number of statistical summaries per entry in the "composition" field. When indexing that structure with UUIDs, we can run into a situation where if there are duplicate UUIDs in the composition field, the index into the result Timeseries structure becomes off.