Comments on: Introducing CUBESET() and CUBERANKEDMEMBER() http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/ Thoughts on PowerPivot, SharePoint BI, and the occasional movie quote Fri, 30 Jul 2010 14:40:21 +0000 hourly 1 http://wordpress.com/ By: JA http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-1059 JA Wed, 31 Mar 2010 10:22:20 +0000 http://powerpivotpro.wordpress.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-1059 Brilliant post and a brillian comment geb - I was having the same problem. Brilliant post and a brillian comment geb – I was having the same problem.

]]> By: powerpivotpro http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-690 powerpivotpro Fri, 12 Feb 2010 01:25:48 +0000 http://powerpivotpro.wordpress.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-690 Geb/Colin/Dany - interested in posting your mods/workarounds? If you don't want to write a full post, you can just upload the workbook and then we just post a few highlights and a link to the file. I've started a Samples library here: http://powerpivotfaq.com/PowerPivot%20Samples/Forms/AllItems.aspx Whatcha think? Geb/Colin/Dany – interested in posting your mods/workarounds?

If you don’t want to write a full post, you can just upload the workbook and then we just post a few highlights and a link to the file.

I’ve started a Samples library here:

http://powerpivotfaq.com/PowerPivot%20Samples/Forms/AllItems.aspx

Whatcha think?

]]> By: geb http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-613 geb Fri, 05 Feb 2010 15:42:44 +0000 http://powerpivotpro.wordpress.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-613 I think I found what I was looking for, maybe I phrased my question incorrectly. Changing the cubemember tuple in the last argument in the cubeset only changes the top sort, (very useful in its own right), what I was looking for was how to limit the returning members of the cubeset based on a parameter in a different cell. If I choose a team in the tuple, the teams player will come out first, but player on other teams will still be in the cubeset. This below is what I came up with after playing around. The cubeset now switches dynamically to list the players on team specified in cell e7, in ranked order of their rushing yards. CUBEMEMBER("sandbox","[team].[" &E7&"]",+e7) (e7 is a cell where you pick a team from a validation list/drop down box/slicer) =CUBESET("sandbox", "([team].[" & $e$7 &"]*[CleanPlayers].[FullName].children)",$e$7 & " top rushers",2, "[Measures].[Rushing Yards]") I couldn't test with Robs actual cube (but he didn't send Kasper the data either), but was able to get it working quite nicely on one of my own. With a few other tricks, we can use this to create dynamically sizing drop down lists using the validation function, and no code. If I can get something cleaned up I can send rob a copy along to illustrate. I think I found what I was looking for, maybe I phrased my question incorrectly.

Changing the cubemember tuple in the last argument in the cubeset only changes the top sort, (very useful in its own right), what I was looking for was how to limit the returning members of the cubeset based on a parameter in a different cell.

If I choose a team in the tuple, the teams player will come out first, but player on other teams will still be in the cubeset.

This below is what I came up with after playing around. The cubeset now switches dynamically to list the players on team specified in cell e7, in ranked order of their rushing yards.

CUBEMEMBER(“sandbox”,”[team].[" &E7&"]“,+e7) (e7 is a cell where you pick a team from a validation list/drop down box/slicer)

=CUBESET(“sandbox”, “([team].[" & $e$7 &"]*[CleanPlayers].[FullName].children)”,$e$7 & ” top rushers”,2,
“[Measures].[Rushing Yards]“)

I couldn’t test with Robs actual cube (but he didn’t send Kasper the data either), but was able to get it working quite nicely on one of my own.

With a few other tricks, we can use this to create dynamically sizing drop down lists using the validation function, and no code. If I can get something cleaned up I can send rob a copy along to illustrate.

]]> By: Dany Hoter http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-612 Dany Hoter Fri, 05 Feb 2010 10:29:17 +0000 http://powerpivotpro.wordpress.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-612 After such an introduction from Rob I really don't have a choice but become active on this site. Both Kasper and Geb questions can be answered by the same technique. When creating a set with CUBESET we don't need to limit the number of members in the set. The important thing is to get the order of members in the set be by descending order of the measure including the slicers. Once we achieve that we can decide how many CUBERANKEDMEMBER functions to use from the top of the set and in this way achieve the effect of a topcount. In order to get the right order we need to define something called a tuple in MDX terminology. An example of a tuple is sales of milk in April/2010. A tuple can contain a single member from any number of dimensions. We can create a tuple using CUBEMEMBER with a range of cells as the second argument. Each cell in the range should contain a member. Assume that A3 contain the measure, B1 the sliced time period and B3 the sliced product category. The formula CUBEMEMBER("sandbox",(A3,B1,B3),"This is a tuple") can be used as the last argument for the CUBESET function (Of course A1:A4 is a valid range in case the cells are consecutive). The result will be a set ordered by the measure value in the filtered period and the filter category. There are still some issues in the solution the most acute one is how to sort if one or more filters use mutiple members. I've send Rob an example and he'll adapt his spreadsheet and publish a post with more details. After such an introduction from Rob I really don’t have a choice but become active on this site.
Both Kasper and Geb questions can be answered by the same technique.
When creating a set with CUBESET we don’t need to limit the number of members in the set. The important thing is to get the order of members in the set be by descending order of the measure including the slicers. Once we achieve that we can decide how many CUBERANKEDMEMBER functions to use from the top of the set and in this way achieve the effect of a topcount.
In order to get the right order we need to define something called a tuple in MDX terminology. An example of a tuple is sales of milk in April/2010. A tuple can contain a single member from any number of dimensions. We can create a tuple using CUBEMEMBER with a range of cells as the second argument. Each cell in the range should contain a member.
Assume that A3 contain the measure, B1 the sliced time period and B3 the sliced product category. The formula CUBEMEMBER(“sandbox”,(A3,B1,B3),”This is a tuple”) can be used as the last argument for the CUBESET function (Of course A1:A4 is a valid range in case the cells are consecutive). The result will be a set ordered by the measure value in the filtered period and the filter category.
There are still some issues in the solution the most acute one is how to sort if one or more filters use mutiple members.
I’ve send Rob an example and he’ll adapt his spreadsheet and publish a post with more details.

]]> By: powerpivotpro http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-608 powerpivotpro Thu, 04 Feb 2010 13:25:24 +0000 http://powerpivotpro.wordpress.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-608 Kasper and Geb - excellent points and questions. I see this as an awesome opportunity to bring in a man who I consider to be the world's foremost expert on cube formulas, the great Dany Hoter. Let me see if I can lure him into this conversation with us. He is very busy, but an in-depth cube formula chat like this is damn near irresistible to him, so let's keep our fingers crossed. If I can't get him to weigh in, I'll dive in myself. -Rob Kasper and Geb – excellent points and questions. I see this as an awesome opportunity to bring in a man who I consider to be the world’s foremost expert on cube formulas, the great Dany Hoter.

Let me see if I can lure him into this conversation with us. He is very busy, but an in-depth cube formula chat like this is damn near irresistible to him, so let’s keep our fingers crossed.

If I can’t get him to weigh in, I’ll dive in myself.

-Rob

]]> By: Cube Formulas and NamedSets Running Commentary « PowerPivot Twins! http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-606 Cube Formulas and NamedSets Running Commentary « PowerPivot Twins! Thu, 04 Feb 2010 09:57:35 +0000 http://powerpivotpro.wordpress.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-606 [...] Introducing CUBESET() and CUBERANKEDMEMBER() [...] [...] Introducing CUBESET() and CUBERANKEDMEMBER() [...]

]]> By: Pivots, Cube Formulas, Named Sets … Oh my! « Denny Lee http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-604 Pivots, Cube Formulas, Named Sets … Oh my! « Denny Lee Thu, 04 Feb 2010 09:44:25 +0000 http://powerpivotpro.wordpress.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-604 [...] out for diving into the deep end into his world of Excel and getting the proverbial egg on my face (Joker picture not-withstanding).  He’s called me out, challenged my geek hood, and sliced / diced my postings (pun intended).  [...] [...] out for diving into the deep end into his world of Excel and getting the proverbial egg on my face (Joker picture not-withstanding).  He’s called me out, challenged my geek hood, and sliced / diced my postings (pun intended).  [...]

]]> By: geb http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-600 geb Thu, 04 Feb 2010 03:55:49 +0000 http://powerpivotpro.wordpress.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-600 What about using the cubeset function with a .children and a limiter? The syntax to brutal for me after trying out a couple times to figure out. eg cubeset top 10 rushers, in week X or top 10 rushers where reciving yards greater than Y something like multiple select boxes, Select the team, and it gives you the list of players in the team by restricting the cubeset2 (player list), using team 1 as an input? or is that what the last sentance of kasper's coment meant,that not being able to use a filtered cubeset is a big disadvantage What about using the cubeset function with a .children and a limiter? The syntax to brutal for me after trying out a couple times to figure out.

eg cubeset top 10 rushers, in week X
or top 10 rushers where reciving yards greater than Y

something like multiple select boxes, Select the team, and it gives you the list of players in the team by restricting the cubeset2 (player list), using team 1 as an input?

or is that what the last sentance of kasper’s coment meant,that not being able to use a filtered cubeset is a big disadvantage

]]> By: Kasper de Jonge http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-594 Kasper de Jonge Wed, 03 Feb 2010 21:13:14 +0000 http://powerpivotpro.wordpress.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-594 Hmm Rob I've been playing with the these formula's and inspected your workbook. Inventive stuff tying it all together like this. If i understand correct you get the 50 players with the most rushing yards (regardless of the slicers). you get the fantasy points from these players (with slicers). And then select the top 20 fantasy points by largest. Lookup the playersname by using the points. And put these two fields at the report. A few remarks: the cubeset and cuberankedmember is regardless of the slicers, isn't it possible to make it work with the slicers? so when you use the slicers the sorting will be applied on the filtered set. The cubevalue is applied with filtering. The way you fix this is by getting a set, use filtered values to get the top n values and then lookup the members belonging to these values. I know i shouldn't think as a developer but this is very unnatural to me. What is the chance a player from a specific weeknumber is not in your top 50 list? Using these functions has great potential but you have to have a specific purpose and know what you are doing. I'll have to play some more with it, currently i see the not being able to use a filtered cubeset and cuberankedmember as a big disadvantage. To be continued :) Kasper Hmm Rob I’ve been playing with the these formula’s and inspected your workbook.
Inventive stuff tying it all together like this.

If i understand correct you get the 50 players with the most rushing yards (regardless of the slicers). you get the fantasy points from these players (with slicers). And then select the top 20 fantasy points by largest. Lookup the playersname by using the points. And put these two fields at the report.

A few remarks: the cubeset and cuberankedmember is regardless of the slicers, isn’t it possible to make it work with the slicers? so when you use the slicers the sorting will be applied on the filtered set. The cubevalue is applied with filtering.

The way you fix this is by getting a set, use filtered values to get the top n values and then lookup the members belonging to these values. I know i shouldn’t think as a developer but this is very unnatural to me. What is the chance a player from a specific weeknumber is not in your top 50 list?

Using these functions has great potential but you have to have a specific purpose and know what you are doing. I’ll have to play some more with it, currently i see the not being able to use a filtered cubeset and cuberankedmember as a big disadvantage. To be continued :)

Kasper

]]> By: kasper de jonge` http://powerpivotpro.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-593 kasper de jonge` Wed, 03 Feb 2010 14:46:21 +0000 http://powerpivotpro.wordpress.com/2010/02/02/introducing-cubeset-and-cuberankedmember/#comment-593 Thank god i don't look like that :) But you're right i had to know and i can't stand teasers. Great stuff these Excel tricks, all new for me as a BI developer. So keep on showing these tips and tricks, maybe i'll create a screencast soon how to create a report like this. What i would like to do is create a filter on the PowerPivot data that i can control with webpart connections! Thus far couldn't get it to work, keep on playing :) BTW what will the Cube be named in RTM is not Sandbox? Thank god i don’t look like that :) But you’re right i had to know and i can’t stand teasers. Great stuff these Excel tricks, all new for me as a BI developer. So keep on showing these tips and tricks, maybe i’ll create a screencast soon how to create a report like this.

What i would like to do is create a filter on the PowerPivot data that i can control with webpart connections! Thus far couldn’t get it to work, keep on playing :)

BTW what will the Cube be named in RTM is not Sandbox?

]]>