![]() | |
#1
| |||
| |||
|
|
Hello, If I already have my data warehouse in SQL Server 2000 and dimensionally modeled into a (star schema), then why do I need Essbase to use Hyperion as a front end for my reports and anlysis? Can Hyperion use my already available structure directly? And why Essbase is different from other RDBMS(s)? Thank you. |
#2
| |||
| |||
|
|
Hello, If I already have my data warehouse in SQL Server 2000 and dimensionally modeled into a (star schema), then why do I need Essbase to use Hyperion as a front end for my reports and anlysis? Can Hyperion use my already available structure directly? And why Essbase is different from other RDBMS(s)? Thank you. |
#3
| |||
| |||
|
#4
| |||
| |||
|
#5
| |||
| |||
|
|
"John Thompson" <jcattm (AT) hotmail (DOT) com> wrote in message news:2f90defc.0308141336.69a14b50 (AT) posting (DOT) google.com... I would think with 77,000 products and around 38,000 customers, you would have to be (excuse me) crazy to try OLAP. UNLESS, you can easily make everything unique. With a cube, you cannot have two of anything the same, including cross dimensions. For example, if you have May as a month, but May is also a customer, or procuct, you have to qualify one as, say, May - Month. Making loading data nuts! Basically, it takes a team of people or 2 years for 1. You have to learn how to think in data and outline. (An outline is how the data "sums" up). If you then change anything at some level, all the data "above" has to be re-calculated. It basically stores every possible combination of data, for instant access. HUGE! Yes, it's fast and the end users love it, but if you are alone (like I was) I wouldn't touch it with a 3 mile pole. Try to use SQL Server as you have it and it can only increase your knowledge. When you say it would be "crazy to try OLAP", presumably you mean Essbase in particular, rather than OLAP in general? For example, comments like, "With a cube, you cannot have two of anything the same, including cross dimensions" do apply to Essbase, but not to other OLAP products. Specifically, Analysis Services has no such restriction, and can easily handle cubes with larger dimension products than those. Also, no modern OLAP stores every combination of data -- that was an old idea that caused dreadful database explosion, so all modern products offer options for pre-calculating some aggregates in advance and doing other calculations on-the-fly. Essbase does still pre-calculate more than most OLAPs, and a lot more than Analysis Services, however, so Analysis Services cubes are always much smaller and process much faster. than Essbase cubes But Essbase compensates by tending to deliver better query performance, though it does depend a lot on the app and the deployment architecture. Nigel Pendse OLAP Solutions http://www.olapreport.com |
#6
| |||
| |||
|
#7
| |||
| |||
|
#8
| |||
| |||
|
|
It sounds like you went about your Essbase efforts the exact way that most people recommend you don't do it. Not educating yourself on the product & its capabilities, as well as data management principles and skills will always leave you in the dark. One of the great parts about Essbase is that if something changes at a low level in the database, you can selectively reload & calc just the affected portions, which would alleviate your calc time pain. Also, there are ways of structuring your database that will significantly bring your calc & query times down, which you may not have been aware of. There are a bunch of other items I would bring up, but I don't think they would be constructive. I am in the same camp as you, agreeing that Analysis Services is a much better option, especially given that they already have SQL Server in house. I do not agree that it will be any easier for them if they don't go through the appropriate implementation steps, as they will constantly hit roadblocks & pitfalls along the way. Sorry for the little rant here, but I don't think you should be slamming a product without having full knowledge of it. However, I do agree with your conclustion. E |
![]() |
| Thread Tools | |
| Display Modes | |
| |