Comments on: Dynamic TopN Reports via Slicers, Part 2

By: Willem

Willem — Fri, 23 Nov 2012 09:23:45 +0000

Just when you thought it couldn’t get any simpler…
Created a measure [remove blank]=((column1+column2+…)=0,0,1) filtered the value settings on the pivot to [remove blank] greater 0 and it works!

You have no idea how happy this has made me! Thanks for the tip and looking forward to many more great posts on this wonderful tool called PowerPivot!!!

By: powerpivotpro

powerpivotpro — Fri, 23 Nov 2012 01:36:35 +0000

I haven’t thought about it much yet, but maybe combine with an IFFER-BLANKER?

http://www.powerpivotpro.com/2011/09/the-incredible-iffer-blanker-filter-measure/

By: Willem

Willem — Thu, 22 Nov 2012 14:18:35 +0000

Thanks for this great post which has helped me in creating a wonderful TopN report, however, I’ve hit a snag and am wondering if somebody knows a solution for this:
I have a master list with ConsumerName including consumers who’ve been inactive in years. When I do a split by a region which, let’s say, has 7 active Consumers and 50 inactive consumers, and select the Top 10 to be displayed, it will display all 57 Consumers as the 50 inactive Consumers are all ranked #8 and that falls within the Top10. Is there a simple solution for this?

By: powerpivotpro

powerpivotpro — Fri, 07 Sep 2012 16:51:55 +0000

I like it

By: Carsten

Carsten — Fri, 07 Sep 2012 15:58:50 +0000

I created a branched measure to switch between two different types of sales. What I did not expect was that I still could use the measure in a cube function without any slicers. By changing the ID Value in the cube function below I could change the way the measure “Sales” was calculated. Mabe it is all clear to you, but I find this fascinating.

CUBEVALUE(“PowerPivot Data”;”[Measures].[Sales]“;”[Date].[Year].[2011]“;”[SalesTyp].[ID].[2]“)

By: Colin Banfield

Colin Banfield — Sun, 08 Jul 2012 14:53:26 +0000

A long time ago, when I did Percentile II, I agreed to shorten the post in exchange for a separate post on TopN. I should really follow through on that promise!

By: powerpivotpro

powerpivotpro — Fri, 06 Jul 2012 01:06:51 +0000

That’s awesome David

I just noticed TOPN for the first time while doing these two posts. You wanna do a post on this topic? It deserves its own, even if it’s short (or long).

By: David Hager

David Hager — Fri, 06 Jul 2012 00:48:40 +0000

Just so I get it on the record (before ALL of the good ideas have been posted ), here is a formula that returns the sum of the n+1 to 2nth records.

=CALCULATE(SUM([SalesQuantity]),TOPN(n,TOPN(2*n,FACTSALES,[SalesQuantity]),[SalesQuantity],1))

TopN this!

By: powerpivotpro

powerpivotpro — Thu, 05 Jul 2012 18:40:21 +0000

Yeah, I understood why the VALUES(Customer[CustomerKey]) approach returned all 1′s, I just didn’t want to explain it, these posts get long

My point is more that this function is over-flexible at the expense of handling simple common scenarios, and there’s no workbable way (that I’ve found!) to rank based on an ID that isn’t displayed.

Using VALUES() against a field that is not on the pivot works great in other situations, but given the way that RANKX has been implemented, that option has been taken from us.

By: Colin Banfield

Colin Banfield — Thu, 05 Jul 2012 18:32:57 +0000

Rob,

It’s not clear why using CustomerKey in the rank formula is causing a issue, but I did note that there is no problem if CustomerKey is used instead of FullName in the PivotTable (with CustomerKey in RANKX). I think that the best option would be to prefix FullName with CustomerKey. Otherwise, if Willie Xu turns up in the topN, how would we know which Willie Xu we’re looking at?

The formula RANKX(VALUES(Customer[CustomerKey]), [Selected TopN Value], ,0) displays 1 for all ranks because each customer is being ranked with him or herself! We need ALL to rank a customer based on the sales of all customers.