Comments on: What Are Unbalanced Data Providers And Why Should I Care? http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/ ...you are in a twisty maze of passageways, all different... Thu, 09 Sep 2010 09:57:30 -0500 http://wordpress.org/?v=2.8.4 hourly 1 By: Pratik Mondal http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/comment-page-1/#comment-1160 Pratik Mondal Tue, 18 May 2010 09:39:47 +0000 http://www.dagira.com/?p=168#comment-1160 Hi Dave, I am stuck very badly,last few days, with merged dimension issue. I have 3 queries in my report,& all contains 'Client_id' as a dimension object & also a separate prompt object 'Client_id'[actually, the report should run for a specific 'client_id' value]. In the report,i created a section based upon 'Client_id' that is also the merged dimension & drill down is enabled on the 'Client_id. Now, the problem is when i refresh all the 3 queries,it looks like below: Client_id : BB002874 col1 col2 col3 col4 |-------|-----------|--------|-------| |-------|-----------|--------|-------| |-------|-----------|--------|-------| |-------|-----------|--------|-------| The block should produce data for client_id=BB002874 & as per universe hierarchy [there are several group under the client BB002874] if i click on 'BB002874', drill down up to group level should be shown. now problem is drill down is not working though drill down is enabled upon 'Client_id'. pls suggest. Thanks, Pratik. Hi Dave,
I am stuck very badly,last few days, with merged dimension issue.

I have 3 queries in my report,& all contains ‘Client_id’ as a dimension object & also a separate prompt object ‘Client_id’[actually, the report should run for a specific 'client_id' value].

In the report,i created a section based upon ‘Client_id’ that is also the merged dimension & drill down is enabled on the ‘Client_id.

Now, the problem is when i refresh all the 3 queries,it looks like below:

Client_id : BB002874

col1 col2 col3 col4
|——-|———–|——–|——-|
|——-|———–|——–|——-|
|——-|———–|——–|——-|
|——-|———–|——–|——-|

The block should produce data for client_id=BB002874 & as per universe hierarchy [there are several group under the client BB002874] if i click on ‘BB002874′, drill down up to group level should be shown.

now problem is drill down is not working though drill down is enabled upon ‘Client_id’.

pls suggest.

Thanks,
Pratik.

]]> By: Dave Rathbun http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/comment-page-1/#comment-1035 Dave Rathbun Thu, 11 Mar 2010 14:20:43 +0000 http://www.dagira.com/?p=168#comment-1035 N_Kumar, you can't "fool" the product like that. There are clearly more than one Resort value for each Year, so the #MULTIVALUE error would be correct. If the values are properly details, then they should be details at the universe level. N_Kumar, you can’t “fool” the product like that. There are clearly more than one Resort value for each Year, so the #MULTIVALUE error would be correct. If the values are properly details, then they should be details at the universe level.

]]> By: N_Kumar http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/comment-page-1/#comment-1034 N_Kumar Thu, 11 Mar 2010 10:26:45 +0000 http://www.dagira.com/?p=168#comment-1034 HI Dave, Greate Post. In Xi R2 Can we include all dim & measure in single block.If yes then how We have requirement like that Ex-Year|Resort|Region|Revenue|no of guest I have created Region & resort as a detail based on merged year it is giving all measure value perfectly but Resort is showing #Multivalue error HI Dave,
Greate Post.
In Xi R2 Can we include all dim & measure in single block.If yes then how
We have requirement like that
Ex-Year|Resort|Region|Revenue|no of guest
I have created Region & resort as a detail based on merged year it is giving all measure value perfectly but Resort is showing #Multivalue error

]]> By: Sriraman http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/comment-page-1/#comment-805 Sriraman Mon, 05 Oct 2009 12:21:14 +0000 http://www.dagira.com/?p=168#comment-805 Dave, now i can understand unbalanced data providers but i do face lots of probelem on using multiple data providers and i posted this issue on BOB forum but no reply to all. Ley say am using Union 1 and Union 2. Am calling result set of data provider in one of the union and other am not calling at all. According to my condition data provider wont fetch any data and obviously one of the union should not give any data.Since other union am not calling result set of data provider it has to give the result but i dont find any single row fetched in the cube. Is it any thing i missed in the data provider. Thanks Sriraman Dave, now i can understand unbalanced data providers but i do face lots of probelem on using multiple data providers and i posted this issue on BOB forum but no reply to all.
Ley say am using Union 1 and Union 2.

Am calling result set of data provider in one of the union and other am not calling at all. According to my condition data provider wont fetch any data and obviously one of the union should not give any data.Since other union am not calling result set of data provider it has to give the result but i dont find any single row fetched in the cube.

Is it any thing i missed in the data provider.

Thanks
Sriraman

]]> By: Dave Rathbun http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/comment-page-1/#comment-732 Dave Rathbun Tue, 01 Sep 2009 00:47:22 +0000 http://www.dagira.com/?p=168#comment-732 Chris, in the exact scenario I used here, there is no fix that will return the correct results. The data simply isn't there and can't be manufactured. In some cases where you might have a dimension object for order number and another dimension object for order date, and you know there is exactly one order date per order number, then you can use the technique you suggested. Create a detail variable for Order Date and link it to the Order Number dimension. In this scenario everything works because of the one-to-one relationship between the two pieces of data. Measure objects can be projected to details. About the only thing you can't do with detail objects are drill and rank. You can do a break or sort or filter on a detail object. Chris, in the exact scenario I used here, there is no fix that will return the correct results. The data simply isn’t there and can’t be manufactured.

In some cases where you might have a dimension object for order number and another dimension object for order date, and you know there is exactly one order date per order number, then you can use the technique you suggested. Create a detail variable for Order Date and link it to the Order Number dimension. In this scenario everything works because of the one-to-one relationship between the two pieces of data.

Measure objects can be projected to details. About the only thing you can’t do with detail objects are drill and rank. You can do a break or sort or filter on a detail object.

]]> By: Chris http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/comment-page-1/#comment-730 Chris Mon, 31 Aug 2009 17:08:00 +0000 http://www.dagira.com/?p=168#comment-730 Dave, so what is the best solution to fix the issue? Creating details can be useful as an extra piece of info on the report, but what if you want to project to that level? The measures only project to dimensions. Is the only way to graft the unbalanced dimensions into one provider or the other? The biggest issue I have with details is that once you have the report defined, the data should be available at that level. Meaning, if you add a detail to a report and it is listed as a field it should be available to project as an aggregate. I realize today that the engine is only working with the base, separate Data Providers, but a nice enhancement would be to allow the combined report dataset to be used as the source of the calculation context. Dave, so what is the best solution to fix the issue? Creating details can be useful as an extra piece of info on the report, but what if you want to project to that level? The measures only project to dimensions. Is the only way to graft the unbalanced dimensions into one provider or the other?

The biggest issue I have with details is that once you have the report defined, the data should be available at that level. Meaning, if you add a detail to a report and it is listed as a field it should be available to project as an aggregate. I realize today that the engine is only working with the base, separate Data Providers, but a nice enhancement would be to allow the combined report dataset to be used as the source of the calculation context.

]]> By: Dave Rathbun http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/comment-page-1/#comment-729 Dave Rathbun Sat, 29 Aug 2009 00:19:26 +0000 http://www.dagira.com/?p=168#comment-729 The next post is essentially done. I usually let a post marinate for a few days and go back and review it before publishing it, so expect it out sometime next week. It does a much better job of explaining that MultiCube() would not fix the scenario outlined in this post, and why. The next post is essentially done. I usually let a post marinate for a few days and go back and review it before publishing it, so expect it out sometime next week. It does a much better job of explaining that MultiCube() would not fix the scenario outlined in this post, and why.

]]> By: Dave Rathbun http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/comment-page-1/#comment-727 Dave Rathbun Fri, 28 Aug 2009 02:40:23 +0000 http://www.dagira.com/?p=168#comment-727 Marek, I realized tonight that my comment is hard to understand as far as the distinction I was making. So I guess my follow-up post needs to be written sooner rather than later in order to explain what I meant. :) Marek, I realized tonight that my comment is hard to understand as far as the distinction I was making. So I guess my follow-up post needs to be written sooner rather than later in order to explain what I meant. :)

]]> By: Krishnak http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/comment-page-1/#comment-715 Krishnak Wed, 26 Aug 2009 13:09:33 +0000 http://www.dagira.com/?p=168#comment-715 Hi Dave, The information what you posted is, help's me a lot Hi Dave,

The information what you posted is, help’s me a lot

]]> By: Dave Rathbun http://www.dagira.com/2009/08/25/what-are-unbalanced-data-providers-and-why-should-i-care/comment-page-1/#comment-713 Dave Rathbun Tue, 25 Aug 2009 19:37:48 +0000 http://www.dagira.com/?p=168#comment-713 Marek, the MultiCube() function will help if data providers are unbalanced "upwards" but not "downwards" as I have discussed in this post. There is no fix for the situation as I outlined it here because the data simply does not exist. There isn't enough information. With an "upwards" unbalanced situation you can force the data down to the lower level, do the join, then roll it up. That's what MultiCube() does. Also, XI 3.x introduced the ForceMerge() function to provide this functionality in Web Intelligence. I have a draft of a post talking about that which will hopefully be done. Someday. :) Marek, the MultiCube() function will help if data providers are unbalanced “upwards” but not “downwards” as I have discussed in this post. There is no fix for the situation as I outlined it here because the data simply does not exist. There isn’t enough information. With an “upwards” unbalanced situation you can force the data down to the lower level, do the join, then roll it up. That’s what MultiCube() does.

Also, XI 3.x introduced the ForceMerge() function to provide this functionality in Web Intelligence. I have a draft of a post talking about that which will hopefully be done. Someday. :)

]]>