Whether you are a user or implementer of Smart View there comes a time when an upgrade is necessary. Since each user installs the smart view binaries on their workstation/laptop, the question of how to push a system-wide upgrade may arise. Fortunately this functionality is built into Essbase Administration Services.
Read the full story

Sometimes I feel like I am the only evangelist out there for Essbase Studio and I wonder why that is since Essbase Studio is such a great tool.  This is especially true when you look at its ability to connect to a myriad of data source such as OBIEE, Teradata, MySQL, EPMA Dimension Server, netezza, etc.  Below is the top 10 list of why an organization should migrate their existing Essbase Integration Services models to Essbase Studio.  If your team is not an early adopter, run it in parallel.  If your team is scared of change, have them read “Who Moved My Cheese?“.  Otherwise, let’s make it happen.  Get with the program and enjoy the fruits of this fantastic toolset made for those who believe analytics are always evolving for the better.

Top 10 List

10. It has a familiar feel to EIS but its actually intuitive and user friendly.

9. It integrates with the new features of Essbase Server 11x.  Has EIS been touched since System 9x?

8. Introspection

7. There is a step-by-step tutorial on Essbase Studio Server made just for you (here).

6. Security provisioning and roles can be assigned in Shared Services.  Deliniate data modelers from administrators.

5. Essbase Integration Services is about to get sunset and Oracle Support will go bye-bye.

4. It connects to EPMA Dimension Server out-of-the-box. Ever heard of Master Data Management?

3. Essbase Studio Models are deployable with native MaXL Scripting.  Have you tried to automate a EIS deployment?

2. Essbase Studio comes with an Integration Services Catalog Migration Wizard. (hint, hint)

1. Larry Ellison wants you to.  Seriously, Oracle’s BI roadmap is taking Studio to the next frontier, not EIS.  Its a good train…get on it.

ASO does not provide an out-of-the-box version of dynamic-time-series (DTS) like its partner in crime, BSO. However, one straightforward approach to achieving this functionality is to implement a time aggregation dimension and leverage MDX functionality.  In this example we will house the Time Periods/Months (i.e.:  Mar, Apr, May, etc.) and the Years dimensions into two separate dimensions.  We then add an additional dimension called “Time Series” to the outline to support our ASO version of DTS.

Read the full story

By now every knows that the release of Hyperion 11x has provided us with the latest and greatest Essbase modeling application, Essbase Studio.  A lot of us are familiar with Essbase Integration Services and have been looking at Essbase Studio like a two-headed dog.  They either are scared and don’t want to go near it or they are so used to their purse-dog, i.e: EIS, that they stick with what they know best and haven’t unwrapped Essbase Studio yet.

This blog post should put and end to any reluctancy.  I introduce to you the first Essbase Studio tutorial on the web. 

You can download the PDF here. (I’ll eventually move this to a CDN if the downloads get to big, so please link to this post and not the document itself.)

Hyperion Essbase Studio Tutorial by Example (PDF)

I am also breaking the tutorial into a series with this being part 1.  There are a lot of cool new functions with Essbase Studio and you will immediately see the benefits it has over EIS. Soon, you’ll see other blog post here on Essbase Studio tutorials part 2, 3, etc. so be sure to check back often or subscribe to the RSS.

If this series is good, bad, or indifferent please let me know.  I would like to keep these up and with your feedback I can revise and make the documents that much more informative everyone.

Cheers.

Update 8/30/2009

- This should have been posted to the live site last week. I am in the middle of switching hosting providers for my blog and uploaded this post to my test site instead of this one. Sorry for the delay.

I wanted to revisit a Hyperion Essbase adage that every Hyperion Essbase Administrator should know by heart - FIX on Sparse, IF on Dense.

If you don’t know this mantra then burn it into your memory and read on to get the content behind it.

When working with calculation (Calc) script in the BSO world, ideally if you are working with allocations or other logical calculations where you need to isolate a smaller set of blocks you will want to FIX on the dimnesions that are sparse. The long and the short of this is that this is really just a best practice. You could of course FIX on a dense dimension as the Essbase parsing engine will allow it. However, doing so will cause you a very inefficient performance hit because multiple passes through the same blocks will take place. On a small cube you probably wouldn’t notice it but a large cube you will take a hit.

Now putting the IF on dense dimensions is again part of the best practice. Doing this will allow the IF logic to be performed while the block is in memory which means it will end up being a single pass operation.

I have to thank Greg V. for bringing this topic up in a recent conversation which led me to this quick post.

BSO still rocks and is alive and kicking in many organizations. And, I’m not just talking about Hyperion Essbase versions < 7 or Hyperion Planning cubes. Even though ASO reared its head in version 7+ an its agg time gets crazy performance gains over BSO we cannot simply move to the new Aggregate Storage option and forget the finer points of BSO and how it all started off in the first place. Keep living BSO, you rock.

We’ve been working with Oracle for sometime now with the latest release Oracle EPM release and there has been a lot of demand for some demonstration and quick training on Hyperion Essbase Studio.  Since this is new to the 11x release there is not a lot for the general developer or user.

Check out the link to a quick 45 minute training video that was put together.  It should help you to get ramped up on the topic.

http://ilearning.oracle.com/ilearn/en/learner/jsp/rco_details_find.jsp?srchfor=null&rcoid=535889535

I plan to provide several tutorials on Essbase Studio over the next few months furthering this demo so check back or subscribe to the RSS feed.

Cheers,

Christian

This is just another post that I’ve been meaning to put up for a while but it just became relevant again on an OBIEE project I am working on.  In Essbase I am generally a fan of loading data in using Parent/Child builds rules but as we know with OBIEE when looking at the relational datasets a parent/child table can prove to be an annoyance.  And, thus the need to flatten a parent-child table into a more OBIEE-friendly horizontal table. Basically how do we get a table schema like:

CREATE TABLE olap.vDimAccounts(
parent nvarchar(50)
, child nvarchar(50)
)

to look like

CREATE TABLE olap.ParentChildFlattened(
level1 nvarchar(50)
,level2 nvarchar(50)
,level3 nvarchar(50)
,level4 nvarchar(50)
,level5 nvarchar(50)
,level6 nvarchar(50)
...
)

This can of course be accomplished using several different techniques.  I am also contemplating writing a simple C# console application to handle this in a more robust command-line interface manner but that can wait.  On this project I have an Accounts dimension. It  has ragged hierarchies, of course.  Here’s my solution:

USE MyDataBaseName
GO

SELECT
	lev01.child level_01,
	lev02.child level_02,
	lev03.child level_03,
	lev04.child level_04,
	lev05.child level_05,
	lev06.child level_06,
	lev07.child level_07,
	lev08.child level_08,
	lev09.child level_09,
	lev10.child level_10,
	lev11.child level_11,
	lev12.child level_12
INTO
	olap.dimAccountsFlattened
FROM
	olap.vDimAccounts lev01
	LEFT OUTER JOIN olap.vDimAccounts lev02 ON lev01.child = lev02.parent
	LEFT OUTER JOIN olap.vDimAccounts lev03 ON lev02.child = lev03.parent
	LEFT OUTER JOIN olap.vDimAccounts lev04 ON lev03.child = lev04.parent
	LEFT OUTER JOIN olap.vDimAccounts lev05 ON lev04.child = lev05.parent
	LEFT OUTER JOIN olap.vDimAccounts lev06 ON lev05.child = lev06.parent
	LEFT OUTER JOIN olap.vDimAccounts lev07 ON lev06.child = lev07.parent
	LEFT OUTER JOIN olap.vDimAccounts lev08 ON lev07.child = lev08.parent
	LEFT OUTER JOIN olap.vDimAccounts lev09 ON lev08.child = lev09.parent
	LEFT OUTER JOIN olap.vDimAccounts lev10 ON lev09.child = lev10.parent
	LEFT OUTER JOIN olap.vDimAccounts lev11 ON lev10.child = lev11.parent
	LEFT OUTER JOIN olap.vDimAccounts lev12 ON lev11.child = lev12.parent
WHERE
	lev01.parent = 'Accounts'-- IS NULL

How does it work?
I can tell from essbase (or an import of my essbase cube into OBIEE) that I have 12 generations so I know that I should have at least 11 columns in my final horizontally flattened table.
Next I create a left outer join on the same table or view object containing my parent-child hierarchy each time referencing the previous instance of said table so that I capture the true parent/child relationship.

The final where clause really just tries to determine where to kick this puppy off. In most parent/child hierarchy tables the first parent is Null to indicate that the first records child does not have a parent. I would do that for a parent/child table when building a MS Analysis Services cube or just a relational report but in Essbase we typically fill that first parent with the name of the dimension if doing a full build so the load rule knows where to start the hieararchial build.

What other techniques exist?
Well this topic is actually somewhat obfuscated since few implementations require it and it is somewhat of a technical work-around. Here are few worthy links on twists on this adaptation:

• Ryan Farley on .Net…
• StackOverflow.com Forum
• Janne Pyykkö

Props On This Knowledge
I have to give mad props to Janne P. on this one because originally his post a year or so ago really helped me out. I thought I’d share the knowledge.

I am delighted today to follow up on my first article regarding on installing Hyperion System 11x on a Windows Server 2003 Virtual Image (VMWare) of Hyperion System 11x. Just to recap, the purpose of this exercise is a “let’s see what will happen” approach to setting up a handy environment to pass on to colleagues and to do some sandbox testing of System 11 without breaking a more permanent infrastructure like the dev boxes we have at our office. Also, I wanted an environment that was minimal. I want Essbase, Shared Services, and sooner or later Hyperion Planning but that’s about it. I am getting ready to work on a cool Essbase / OBIEE integration as well as a Hyperion Application Builder.NET. Anyway…

So as you may recall, we were also using this exercise to compare two documents that bubbled up last year by Oracle and Applied Olap (good ‘ol Tim) to see which one was worthy of your time. Those documents can be found here:

• Oracle Hyperion 11x Installation Guide
• Applied OLAP’s 11x Installation Guide (alternatively in PDF)

Cutting to the chase, both of the documents are very relevant to the exercise at hand.  The Oracle document provides what I would consider a full implementation installation while the Applied OLAP (AO) document is a more concise “get you up and running” overview.  The AO document does a much better job of explaining caveats and the rationale behind why you put a check in a checkbox rather that leaving it unchecked.  The Oracle document does not take that into consideration at all.

I must have had a later release of the 11x installations as my prompts were slightly different from the screenshots on either document in some places. An example of this is the “Set the ARBORPATH Environment variable” option on the essbase configuration prompt.  This was not visible anywhere during my installation.

arbor path env var option

Ultimately for this exercise AO’s document was exactly what I needed. I used the Oracle document as a sub-guide just to make sure I was adhering to Oracle’s standards and not just AO’s.

Instead of an Oracle database I switched up and used a SQL Server 2005 installation.  I bumped into an error with port 1433 (default instance listener port) but was able to resolve that with a setting of firewall configuration.

Post-Installation Problems

But of course, as with anything in IT, these installations do not come without incident.  I encountered the following issues post-installation and will discuss them with some resolutions in the next part of this series:

• “Unable to Bind Host Server Socket On Port [1423]“  error trying to start up essbase following the AO document
  • I came across this error when attempting to run the “essbase.exe” executable from the command line ast Tim stated in his blog post.  Luckily one of his faithful followers assisted and offered the following comment:
    

    “But in order to run essbase in the last step from ‘products\Essbase\Essbase Server\bin’, you first have to stop the Service name “Hyperion Essbase Services”, or else you will get an error message saying “Unable to Bind Host Server Socket On Port [1423]” (network error 10048.”

• AS service not started after install
  • After I conducted the install the AS service was not up and running.  This was not a big deal and I started it manually just to see if the install was completely functional.  When I rebooted the VM Image all necessary Hyperion services were online and good to go.

So we at The Art of BI decided it was time to stand-up a sandbox version of Oracle Hyperion System 11 (11.1.1) since we were tired of using our client’s boxes for research, training, fun development stuff, etc.  : )

 The idea here is to get a nice single server instance of the EPM suite up with the core Shared Services, EPM, Essbase, and Planning.  In so many ways we are doing it both to see if it can be done and what will happen when we do. Other questions we’ll ask is How does it perform? What are the limitations? And, What are our disk space and memory issues, If any?

By the way we will do this on a Parallels VM using Windows Server 2003 Enterprise Addition.  Should we use MySQL now that they are under the Oracle umbrella? We’ll see about that one. : )

To aid our installations we are going to reference both of the following installation docs and contrast and compare the two:

• from Applied OLAP
• from Oracle Technology Network

We’ll provide some feedback and maybe some screenshots during the next couple of days and let you know how it went and what our findings were.

See Part 2 of this series >>

We’ve been doing quite a bit of Hyperion Essbase integration into Oracle BIEE lately.  It is just cool stuff and most Oracle BI consulting groups will admit that this is all the rave right now.

Although our group has published several documents internally on the subject and there are several bloggers who discuss the topic I just the other day found Oracle’s best practice guide to the integration.

It is a short and sweet PDF.  In my opinion it could be better (always room for improvement).  Maybe I’ll update the document in my copious time : )

Here’s the link to the PDF:

Oracle’s Best Practices Guide for Implementing Essbase as a Data Source for OBIEE (PDF)