1 of 1 people found this helpful
Could you please change the aggregation
from COUNT([Consecutive Number])
to COUNTD([Consecutive Number])
and see if results would be closer to expected?
I suppose that even if your accident table
contains one row per unique Consecutive Number,
your other tables (vehicle and person) may contain
more than one row per unique Consecutive Number.
That is if more than one vehicle (or person) is involved in accident.
That could lead to a finer granularity of your resulting dataset,
namely one row per unique accident - vehicle - person combination.
If that's the case, then you're better to use
distinct counts on some of your Measures,
or better to use Level-Of-Detail (LOD) calculations.
Hope it could help.