How to do sequential filtering for rankings

[LucaixBolha]

Hi guys! 

I have a dimension table of Assets, and I'm doing a three-step ranking filtering, being that after each rank filter the number of assets analyzed is restrained.

I created independent slicers for each ranking (Market Value, Volatility and "P/BV", known as Price/Book-to-Value), but when I apply the third ranking it filters the Assets of the first rank's selection (Mkt Value):

I would like to restrain the Assets range to the codes in the volatility ranking (AESB3, ABCB4 and TTEN3), and then apply the P/BV rank. The expected result would be this:

This is the link to the example pbix: https://drive.google.com/file/d/1N_g1fl5zkXibp6vNwy9s_oBifJCiGu0n/view?usp=sharing

Could anyone give me an advice on this problem please? 

Thanks!

[LucaixBolha]

Didn't manage yet to do it 😬, anyone willing to help? 😅

[lbendlin]

TopN 3 P/BV = 

VAR SelectedTop = SELECTEDVALUE('TOPN 3'[Top N 3])

RETURN
CALCULATE(
SWITCH(
    TRUE(),
    SelectedTop = 0, 
    [_Rank Switch P/BV],
    
    [_Rank Switch P/BV] <= SelectedTop,
    [_Rank Switch P/BV]
),

SelectedTop will never be zero. You would need to use the standard measure trick

VAR SelectedTop = 0+SELECTEDVALUE('TOPN 3'[Top N 3])

Adjust your other measures accordingly.  Note:  You may need to compute your TOPN dynamically because the TOP N 3 filter should only allow you to select whatever has been prefiltered in Top N 2.  (In your example Top N 3 should only allow you to select from 1 to 3 items, no?) 

[LucaixBolha]

@lbendlin, could you please talk a little bit more about the standard measure trick (what is it for, etc.)?

[lbendlin]

If there is a threat that a measure could return BLANK()  you can add a 0 to the measure output to return at least "something"  without skewing the final result. 

[LucaixBolha]

ahhh got it! Thanks a lot 😄

[LucaixBolha]

Ahh you're right @lbendlin!

Indeed, Var SelectedTop won't be zero and TOP N 3 filter should only allow me to select whatever has been prefiltered in Top N 2. Thank you very much for your time, I'll try it out! 🙏🙌