
1. Re: Excluding duplicate cases in aggregate measures
Simon Runc Jul 30, 2016 3:33 AM (in response to Alex Trouteaud)hi Alex,
So the link you've provided is one way, but that was before we have FIXED LoD calculations. Prior to this you would have to do this using Table Calculations. Table Calculations are Aggregates of Aggregates, so we could take the MIN([Number of Records]) for each Fruit, and then WINDOW_SUM that result. The issue here would be that you'd always need Fruit in your VizLoD (i.e. in the Viz) in order for the Table Calc (and the Min) to run correctly.
Now we have FIXED LoDs this task becomes much easier. In essence FIXED works the same way (in this example) but we can define the LoD that the Aggregate is run over, in the formula, and so will work regardless of our VizLoD. You might find this Quora answer on the types of calculation in Tableau and how to think of them useful
So for you issue, I created
[Number of Records  Remove Duplicates]
{FIXED [Fruit]: MIN([Number of Records])}
So this runs a MIN Number of Records against each Fruit (so returns 1 for each Fruit)..and this result is returned at Row Level (at the level of Fruit) which means we can aggregate it again. So I've brought this in as a SUM and I just get a 1 for each fruit.
Hope this makes sense, but please post back if not

2. Re: Excluding duplicate cases in aggregate measures
Alex Trouteaud Jul 30, 2016 5:10 AM (in response to Simon Runc)Thank you Simon  this was very helpful! I modified it slightly to produce what I needed, which was unique values across the [identifier] dimension:
[Number of Records  Remove Duplicates]
{FIXED [identifier]: MIN([Number of Records])}
Worked perfectly...thank you!