Starting services for OBIEE Respository - Business Intelligence Applications

Can making a few minor changes to a Logical BMM layer or the Presentation layer cause the services to not start up again? 


It might in some cases if you have an initblock determining the value of a variable that is used as a fragmentation source. If you initblock fails the model might get invlaidated.
In general look for initblock failures at both repository and session levels. 

If your repository passes a consistency check it should always start.
Initilization blocks if built using best practices will have default settings.


Does it require to Refresh the Physical, Business and presentation layer after any changes in Database view level changes?

Dear Experts, I am new in Oracle BI, i have encounter an issue where one fact table data was not showing in the report. later i have identified the issue which is associated with the FACT Aggr View , now Fact Aggr has been correct in database, but in Obiee its not reflecting.. my question is do  i need to refresh the entire Physical, business and presentation layer or only for the particular issue table/view.if i need to refresh the obiee repository then how i can do (what are steps? i should follow) . any help is highly appreciated! Many Thanks,Bhasker 
Hi,Not sure I got your question right, so here are the 2 scenarios:a) If you are talking about data (tables/views content) in your database, you don't need to refresh anything in the RPD. If your data change OBIEE will automatically show you the changes as it queries the database and doesn't really store data. Only exception to this is if your have cache enabled, in that case you probably hit the cache and not your own database. Look at your logs and you will see if OBIEE is querying the DB or get results from the cache. b) If the structure change, so your tables/views have new columns or totally new tables/views you will need to change the RPD to reflect the changes as required as OBIEE doesn't do anything automatically. The RPD contains the model and therefor it's up to you to adapt the RPD based on the structure changes in your database. So ... pick A) or B) 
If the table already exists in the physical then you should be able to edit the columns changed manually; I find this faster unless you are talking a massive change; and then drag the columns onto their equivalents in the business model layer to effect the change on the BMM. Presentation layer should not need touching. You can query on the physical to the BMM objects linked to it to find everything that needs updating down the 'food chain'.
Gianni beat me to the 'Add Reply' button - I was describing scenario B.
Hi, Thanks for your response! Can you please tell me where to look for the logs and how i will know OBIEE is querying the DB or get results from the cache. Thanks,Bhasker
You can start by looking in the front-end in the Administration page where you see the active sessions you see the various analysis executed and with their logs.
See also the advanced tab in answers, there is a check box so you can ensure that it does not use cache, invaluable during testing!
Hi, i have checked the log and found report is directly querying the DB and also checked the physical layer we are getting the data for the particluer view-"AGGR_FACT_DEP_ACT"  and alos check relationship/ join with other table which also seames fine, i cound not able to understand what's going wroung with it. . is there any way to view the data in BMM and presentation layer to know where is gap or any alternative to reslove this issue.  Many thanksBhasker
its not using useing the cache
So no use keeping StackOverflow up to date I guess?
Bhasker Kumar wrote:  i have checked the log and found report is directly querying the DB  1.) Ok is it a BI Publisher "report"? Or is it an OBI "analysis"? It's two different object types from two different technologies. Be precise. 2.) So what does that statement mean? It is "directly querying the DB"? Is it a direct DB request analysis which just uses the connection pool definition and totally bypasses the RPD modelization? Because then your whole question becomes irrelevant as there is zero dependency on the RPD.
Hi, 1. its not a BI Publisher Report- its an OBI Analaysis Report. 2. "directly querying the DB " mean that report Query doesnt uses any cache, its querying the Presentation layer subject area tables. i have a simple problem that i could not able to get the data from one table- "OBI Analaysis Report" where i have checked in Phyical layer it has the data but i am not shure how to check the data in BMM Layer. Can you please help ?? Regards,Bhasker
Bhasker Kumar wrote: OBI Analaysis Report. No, it's an "analysis". There is no such thing as an "analysis report". The correct wording is important since it allows to distinguish between technologies and also to know what you're talking about. If you tell your doctor your "foot arm" hurts then he won't know what you mean just because both are extremities of your body. And if you keep insisting of your "arm" even though you mean "foot" then he will analyze the WRONG extremity and you will be frustrated and think the doctor is useless even though the problem comes from your utilization of the wrong word. Bhasker Kumar wrote: i have a simple problem that i could not able to get the data from one table- "OBI Analaysis Report" where i have checked in Phyical layer it has the data but i am not shure how to check the data in BMM Layer. You do not "check the data in the BMM layer". Physically - meaning in the physical layer of the RPD - you can check the data by choosing "View Data" on any table object. Logically - meaning in the BMM Layer - you construct the logical model and then expose it through the Subject Areas of the presentation layer. From the front-end, if you push the log level for queries to level 7, you can see the query moving across all layers: presentation, logical, physical. You can see which LTS is chosen and how the actual physical SQL looks like when it is sent to the database. You have to be precise and careful what you look at. Strip the analysis down to its core. Don't keep 50 attribute useless columns  which clutter things. Get the full log - as I said level 7 - and look at it in detail. Post it here and point out exactly where the issue is. Be. Precise! Badly written questions without sufficient information and detail help nobody. Again: think about how you explain a physiological problem to your doctor. "I feel bad." won't help. "I have an ache on my arm" when it's on your "foot" won't help.
Hi Bhasker, Sometimes, in Physical layer, If cache is enabled for table/view it will pick the data from query result. Uncheck the cacheable or cache persistence in physical layer properties of a table. No need to refresh any layer everytime if the structure of the table/view is going to be same. Hope it helps! 
He said 2 days ago that it isnt using cache.

Row Level Security on a Business Model Using both Essbase and Oracle

Has anyone tried and tested implementing the row-level security for a business model which use both Essbase and Oracle database as the data source? So, let's say a US user can only data for US only when seeing the reports in both cube and relational level, etc.
I have tried searching any post here, but only found posts about security implementation for relational database only. Thank you very much for the help. 
IMO what ever the data source is when u set up variables having the group values initialized and if the variables are used in filters for row-level security, it has to work! 
Is the session variable connection pool dependent or not? If the session variable use a connection pool A, can it only be used for business models that use the connection pool A? Or can it be used for every business models regardless the connection pool used? Thanks. 
Hi, any confirmation regarding my last question above? Thanks.

RPD issue

Hi guys,
i have created a view based on base tables in the database and iam creating an RPD but now there is only 1 folder in it,here my doubt is how can i join self join to the 1 folder or logical diagram in business model or physical layer....Is the way iam going is right??
Instead of creating RPD,can i use the view created in database directly in the answers i.e. can i access the view in database directly in answers screen...if so how?
An ideas are appreciated.
If you want to query on the database directly, there is an option called "Direct Database Request" in the Answers screen. But to use that, you need to have a connection pool for that database defined in the physical layer.
And, reg ur first question "1 folder in it,here my doubt is how can i join self join to the 1 folder or logical diagram in business model or physical layer....Is the way iam going is right??" ---> here folder means, the physical display folder ?? or physical Table???
If it is a physical table, and you want only that table to be queried, you can have an Alias of the same table, join both of them and use it in ur BMM.
Hi suma,
thanks for your prompt reply.
Ya i have read about direct database request of answers,but same way i dint understand connection pool setting in the physical layer???...this physical layer your talking is the RPD physical layer or something else.
Ya your guess was right its physical table,but if i take the alias of it dont i get any data inaccuracy problems
yes.. its the physical layer in the rpd that I mentioned.
You wont get any data inaccuracies as its a self join. 
When dealing with answers,whats the need of RPD physical layer and its connection pool.We assumed that in answers we can directly uses the views and setting connection pool in physical layer.Can you be more clearer or explain me in detail 
you can use the Direct DB reqst directly with out buiding any physical layer or BMM layer objects, provided the query is small and no complicated joins are involved....
But OBIEE objectve is to give a "Reporting Solution", you have to have a model defined , so that one model would cater to the needs of all the reoprting requirements.
1 )In most of the cases, OBIEE end users will not be given connection pool properties. (security reasons)
2) Its not mandatory for the end users to have knowledge on SQL to use OBIEE (I state this taking example of the organisation I work with).
If we create the model and give them the basic knowledge of how to make report etc... that is enough for them....

Consistency Checking Perform

what are the minimum criteria to pass consistency checking for a given repository?
All joins should be properly created,All BMM columns should be properly mapped to physical columns..etc ....
This question covers whole OBI in fact.ARe you getting any error while doing consistemcy check.
>what are the minimum criteria to pass consistency checking for a given repository?
Each and every thing that you in rpd should be consistent.I mean joins that you are creating between facts and dimensions in Physical layer.Hierrarchies that you are creating for Dim tables in BMM layer(Setting content levels in Fact LTS,should have primarykey of Dim in its hierarchy detail level) and joins in BMM layer.Many more things..
Check this...Import steps to remember while doing rpd...$28RPD$29+Design.pdf
Edited by: Srikanth Mandadi on Oct 15, 2010 3:45 AM 
hi Sandhya,
If you go thorugh the BI Admin guide CTRL+E-->Options -->help u will find the plenty of information about the best practices we need to follow
Repository metadata must pass a consistency check before you can make the repository available for queries. The Consistency Check Manager allows you to enable and disable rules for consistency checks, navigate to and fix inconsistent objects, and limit the consistency check to specific objects.
The Consistency Check Manager does not check the validity of objects outside the metadata using the connection. It only checks the consistency of the metadata and not any mapping to the physical objects outside the metadata. If the connection is not working or objects have been deleted in the database, the Consistency Check Manager will not report these errors.
The consistency checker returns the following types of messages:
Error. These messages describe errors that need to be fixed. Use the information in the message to correct the inconsistency, and then run the consistency checker again. The following is an example of an error message:
[38082] Type of Hierarchy '"0RT_C41"..."0RT_C41/MDF_BW_Q02"."Product Hierarchy for Material MARA"' in Cube Table '"0RT_C41"..."0RT_C41/MDF_BW_Q02"' needs to be set.
NOTE: If you disable an object and it is inconsistent, you will receive a message asking if you want to make it unavailable for queries.
Warning. These messages indicate conditions that may or may not be errors, depending upon the intent of the Oracle BI Administrator. For example, a warning message about a disabled join may be the result of the Oracle BI Administrator intentionally disabling a join, such as eliminating a circular join condition. The following is an example of a warning message:
[39024] Dimension '"Paint"."MarketDim"' has defined inconsistent values in its levels' property 'Number of elements'.
Best practices. These messages provide information about conditions but do not indicate an inconsistency. The message appears if a condition violates a best practice recommendation. The following is an example of a best practice message:
[89001] The Physical Table '"XLS_Forecast"."Forecast".."Sheet1$"' has no keys defined.
Apart from these joins,content levels,column formulas ,syntax's ....
source Admin guide
there are many things to be considered..
content levels, joins, proper mapping of logical column to physical column, Keys in the Logical dimension tables and hierarchies, your BMM should have corresponding Presentation catalog.
Once all of above are okay then your BMM will be available for queries(folder turned to green color from red)

Multiple Connection Pools for Same Physical Structure with RPD

I am using OBIEE and have multiple physical databases which contain the exact same structure. Instead of creating the same physical structure within the RPD file with their own connection pools I would like to have one physical structure with multiple connection pools. I have this setup, however, I can only see data from one of the physical databases using the right-click, view data... option within the physical side of the RPD file. I have searched and could not find a solution to my particular issue. Can anyone point me in the right direction? 
I believe this is the designed functionality as the data change should be seen at runtime not design time.
The designers probably kept with the logic that since the structure is the same the data is going to be relatively similar. I believe it pulls from the top most connection pool, right? 
Yes, the data appears to be from the top data source connection. However, I cannot see my second connection even within OBIEE Answers. Maybe I am missing something in my setup. 
You can have multiple connection pools within a physical database in the physical layer of the rpd. When you try to view data for one of the objects in that physical layer, it prompts you to select any of the connection that is associated to that physical database for viewing the data. 
Your connection pool selection from the physical layer for Answers is dependant on the permission settings. If you have access to multiple the connection pools, the top most in the list of the ones you have access to will be used for querying the database when you send the query request from Answers. 
Is there a way to see combined data from both data sources? The data contained within both data sources are different. 
Is there a way to see combined data from both data sources? The data contained within both data sources are different.Yes, You can but in that case you have to go for 2 different sources in Physical Layer i.e. one CP in one database object and other CP in new database object....which is nothing but working with Multiple Data Sources..
Refer something like..
This requires a proper data modeling.
Hope your doubt is cleared. 
Thank you for the link. Is there any way to get around having two separate physical data source structures? I can get this to work using two separate physical data source structures and combining them within the Business Model & Mapping layer. However, from a maintenance perspective of adding multiple data sources with the same structure I was trying to get away with creating multiple connection pools on the same physical layer model and having one Business Model & Mapping layer as well. If this is not possible to accommodate within the physical layer I will have abandon the idea. 
Goal is to get the CPs seperated ..and then you can use both data. But better to have 2 seperate objects. 
Thanks, Deepak. I will separate the physical structures since it is not possible to combine two physical data sources within the physical layer using multiple connection pools. 
Can you try these things
Approach 1:
Create 2 data groups (GroupA, GroupB), in connection pools( ConnectionPoolA,ConnectionPoolB) assign ConnectionPoolA to GroupA only, and ConnectionPoolB to GroupB.
Assign both groups to User. I'm not sure whether this will work or not.
Approach 2:
You said structure are same are the dimensions values also same ? If so you can use fragmentation for facts
Approach 3:
Create a materialized view too produce one final structure
My issue with your suggestion for approach 1 is that every user logging into the application is coming from an IDM LDAP store for which I do not have assigned groups that I can assign to the connection pools. Regarding approach 2, I do have dimensions that are the same as the facts (both fact and dimension logical tables are created from the same source table). Regarding approach 3, I need these data sources to be dynamic since they will be coming from OLTP data sources; a materialized view is not an option for me.
Thank you,