Intersection using ONLY filters- issue

[ioannagr]

Hello so I got the following question : Find all the people  who BOTH during 1-4/2020 AND  9-12/2019 placed orders.

My filters look in order like this:

Year  |  Month  |  Person ID    and then I have a   count of Orders KPI.

What i did is :   Select Year: 2020, then  Select Month: Jan, Feb, Mar, Apr 

"Select possible" from Person ID's

Then select  Year: 2019 , then Select Month: Sep, Oct,Nov, Dec. 

The outcome is not logical.

Please help, it's kind of urgent and i always had this issue :<

Thanks a lot

[eliran]

Hi,

The steps you mentiond looks fine to me.

You mention that the outcome doesnt make sense, maybe try to test yourself.

Make a chart with if / set analysis statement with these two period of times, and see what are the results.

Compare the above result to the result you get when you use the selections.

Eliran.

[ioannagr]

Hi Eliran and thank you for the response, what would be the way you would write the above in set analysis?

I would write it as : 

count (distinct {<customerId = P ({<Year={'2019'},Month={'Sep','Oct','Nov','Dec'}>}customerId ) >*<customerId= P ({<Year={'2020'},Month={'Jan','Feb','Mar','Apr'}>} customerId ) > } customerId)

Comparing them, with filters i get 179K  and with set analysis 35K.

The latter is the correct one, but I need my clients to easily use the UI with steps for every intersection they want and this is getting too difficult for such a simple query.

I don't want the union, i want the intersection.