Comments on: Creating useful PowerPivot data models for public consumption via Reporting Services http://performancepointblog.com/2009/11/creating-useful-powerpivot-data-models-for-public-consumption-via-reporting-services/ A Blog about PerformancePoint, SQL Reporting Services and all the nummy BI technologies that interact with them Mon, 08 Mar 2010 19:02:00 +0000 http://wordpress.org/?v=2.9 hourly 1 By: Some recent PowerPivot links « PowerPivotPro http://performancepointblog.com/2009/11/creating-useful-powerpivot-data-models-for-public-consumption-via-reporting-services/comment-page-1/#comment-539 Some recent PowerPivot links « PowerPivotPro Fri, 04 Dec 2009 06:43:13 +0000 http://performancepointblog.com/?p=124#comment-539 [...]      http://performancepointblog.com [...] [...]      http://performancepointblog.com [...]

]]> By: Russell http://performancepointblog.com/2009/11/creating-useful-powerpivot-data-models-for-public-consumption-via-reporting-services/comment-page-1/#comment-462 Russell Fri, 27 Nov 2009 14:44:28 +0000 http://performancepointblog.com/?p=124#comment-462 You're not missing anything from what I can see - if you want SSRS's SSAS designer to work properly against the model, you'd want to add each "measure" as an explicit measure in the model. That could potentially be quite a bit of work. I guess the jury is still out on what percentage of people will want to leverage a PPvt model from Reporting Services, however...so the whole issue could become a moot point in the real-world. It'll interesting to see. Your idea about presenting one list of tables to the user is a good one, but I suspect it would take lots of work by the product team to implement..and frankly I'd rather see them work on other features they've been promising like alerting, etc. Right now, if you drag an attribute (what should be a measure) from a dimension, the designer treats it as an attribute - so you don't get any data back like you would if you drag the same animal from the measure group. In fact, if you do something like drag in your "attribute" (should be a measure), one of those "count of measure name" measures (to get data to return) and some other random (true) dimension attribute, the query designer chokes when you try and exit - I guess it doesn't like the MDX which gets generated What's in the KPI group? Good question - never bothered to loook. I will next time I fire my machines up. Only the laptop is allowed to be on by order of she who must not be named :) You’re not missing anything from what I can see – if you want SSRS’s SSAS designer to work properly against the model, you’d want to add each “measure” as an explicit measure in the model. That could potentially be quite a bit of work. I guess the jury is still out on what percentage of people will want to leverage a PPvt model from Reporting Services, however…so the whole issue could become a moot point in the real-world. It’ll interesting to see.

Your idea about presenting one list of tables to the user is a good one, but I suspect it would take lots of work by the product team to implement..and frankly I’d rather see them work on other features they’ve been promising like alerting, etc.

Right now, if you drag an attribute (what should be a measure) from a dimension, the designer treats it as an attribute – so you don’t get any data back like you would if you drag the same animal from the measure group. In fact, if you do something like drag in your “attribute” (should be a measure), one of those “count of measure name” measures (to get data to return) and some other random (true) dimension attribute, the query designer chokes when you try and exit – I guess it doesn’t like the MDX which gets generated

What’s in the KPI group? Good question – never bothered to loook. I will next time I fire my machines up. Only the laptop is allowed to be on by order of she who must not be named :)

]]> By: Colin Banfield http://performancepointblog.com/2009/11/creating-useful-powerpivot-data-models-for-public-consumption-via-reporting-services/comment-page-1/#comment-460 Colin Banfield Thu, 26 Nov 2009 21:13:32 +0000 http://performancepointblog.com/?p=124#comment-460 Hi Russell, Am I missing something here? What you seem to be suggesting is that every possible measure that the user can view should be turned into a calculated measure, so that the measure can show up in the measures group. You'd have to do this because you can't anticipate what measures the user *might* use, given the available list. Plainly, this isn't a practical option. The solution? Very simple. Just like the PivotTable field list in Excel makes no distinction between dimensions and measures, neither should Report Builder's metadata list (I bet you weren't confused by the PivotTable field display of the *same* measures and dimensions). But since RB sees PP as an AS source, its field display would have to be reconfigured based on recognizing PP as a special AS source. Simply list all of the tables once and the whole measures vs dimensions mess goes away. BTW, what's inside the KPI group? Hi Russell, Am I missing something here? What you seem to be suggesting is that every possible measure that the user can view should be turned into a calculated measure, so that the measure can show up in the measures group. You’d have to do this because you can’t anticipate what measures the user *might* use, given the available list. Plainly, this isn’t a practical option.

The solution? Very simple. Just like the PivotTable field list in Excel makes no distinction between dimensions and measures, neither should Report Builder’s metadata list (I bet you weren’t confused by the PivotTable field display of the *same* measures and dimensions). But since RB sees PP as an AS source, its field display would have to be reconfigured based on recognizing PP as a special AS source. Simply list all of the tables once and the whole measures vs dimensions mess goes away.

BTW, what’s inside the KPI group?

]]>