Yes it is, but as big corporations usually run multiple data base platforms, Business Objects offers the functionality and flexibility to run against those multiple sources.
True, in most instances, less true in the case of data base appliances, Netezza or Terradata being a couple of examples. Also, don’t forget about the potential impact of column databases, such as Sybase IQ.
Also, what about SAP Hana?
Have you seen, or used, any of the BO OLAP offerings?
Mak 1 
(BOB member since 2005-01-06)
With the universe you can truly screw things up but you cant improve base DB performance. The universes I have created so far have been on well designed Kimball dimensional models with appropriate indexing. At the metadata level your job is to create correct joins/direction of joins and aggregation ruile ect…
Consider this question, “What is wrong with our average handle times this week - they are up by 20%?” (length of time a call center agent spends on the phone with you).
With Qlikview we can set a filter to this week and instantly have only “this” weeks data. Or maybe we want to have this weeks and last weeks side by side for comparison. Now with the correct data, you can “select” just the AHT KPI in a view that has agent as a dimension. One click to sort it by AHT descending and at the top are the agent with the longest handle time. CTRL click the agents with the highest handle times and now you have filtered you data on the problem area. Now you can look at other views. Did these same agents perform better last week? Are they new agents (ie they didn’t exist last week). We can flash over to a call type view (still filtered on the two weeks for the problem agents) and see if the calls this week had a disproportionately high percentage of say “support” calls that take a long time.
All of this in a few seconds utilizing all sorts of graphical object types as you might imagine. Its truly a pleasure to use. Note that MS Excel on SSAS pivot tables is even better but not as pretty.
twebber (BOB member since 2009-03-23)
But surely base DB performance is an issue for any BI tool? Somewhere along the line they all need to retrieve the data, no?
In my experience, BO has a pretty good internal optimizer, certainly BO retrieves rows faster for the same piece of SQL in a number of situations.
Interesting how the ‘instantly’ works (in relation to the previous point) I guess there’s still a data retrieval stage somewhere?
Ah, so Qlikview has some predefined filter types I guess? (sorry, only passing knowledge of the product based on a demo version) I guess this is something we’d assume was a job of the universe/developer if people wanted fast access to ‘this week’ and ‘last week’ as a filter object.
The other stuff you describe sounds like the sort of thing I do on a daily basis with BO, just perhaps through different methodologies, and perhaps slightly longer (by the sounds of it) to create the different views, pivots, etc
e.g. Take the example you gave. Table of users by AHT KPI.
In Deski/Webi, filter on the top 3 users for ‘this week’. Copy paste the table, promote the user filter to global/report, then change the date filter on the second table to ‘last week’. You have your view of the 2 weeks, for the same users.
I would agree that BO seems to be moving away from the analysis side of things - a complaint I had with BI4 and the demise of Deski was that what was once (partially) a quick and dirty analysis tool, was now becoming much more targetted at the production reports arena. Certainly I find data analysis slower (not impossible) in Webi than Deski.
norty303 (BOB member since 2003-03-19)
Norty,
Qlikview is an “in memory” solution. However, the cubes data still needs to be loaded at some point. Please note, this has no real time data loading components like Hana, Quartet or Hyperrig. The later two load cubes one after another pushing the mew one in, when the data becomes “stale”. These are normally used against Financial Front Office applications where “real” real time reporting is, actually, required.
Cheers,
Mark.
Mak 1 (BOB member since 2005-01-06)
Thanks Mak, so it’s like a report with saved data in it essentially? Until the next refresh?
On that basis, isn’t BO as much ‘in memory’ as Qlikview?
Apologies for the derailment 
, I know what I know, but don’t necessarilly know what everyone else is calling it!
norty303 (BOB member since 2003-03-19)
The way you do it with Qlikview or MS SSAS is extract the data form the database incrementally and built it into your solution. For Qlikview you might do it nightly and pull “todays” data into your solution. The refresh of the Qlikview solution is then loaded into memory and available for querying. You might make one trip to the database for every table you need to refresh. The solution is then in memory and your queries will be near instant depending on whether the solution is on a virtual machine or physical machine.
There is not predefined filters for queries. Nothing magic there. I suspect you can do lots of this in BO but so far, I cant see how to do it easily so that end users can do this type of analysis for themselves.
I will keep digging to see what BO has to offer in this area.
twebber (BOB member since 2009-03-23)
Ah, I see, so the difference between QlikView and BO is that BO would be a nightly schedule for the full data set, whilst QV would just obtain the last days data and add to the ‘pot’ so to speak?
What about historically changed data, or does that not exist in your world?
norty303 (BOB member since 2003-03-19)
Norty,
Of sorts, you could compare the BO microcube to an OLAP cube.
Loading OLAP os no different to an incremental load in any RDBMS database platform. In the SAP BW (OLAP) world this is called a delta load, for example. There are occassions, however, where the full cube would have to be re-loaded, if for example there was addtional elements that are added that would effect all data.
The main considerations, when it comes to Qlikview limitations, I discussed in my earlier post.
I believe the important thing to realise here, is what twebber has alluded to there is a difference in regular reporting and OLAP analysis. My belief is that you always start with a data warehouse and then build / derive the nice aggregations and OLAP cubes from there.
Cheers,
Mark.
Mak 1 (BOB member since 2005-01-06)
Thanks Mark for the clarification.
norty303 (BOB member since 2003-03-19)
interessting to read this posts:
not very close with BO but in business with it since v4.5 I remember the “most buggy version 6.5” and had some hopes with SAP integration,
and I found some helpfiul lines…
But I had to notice all the things allready mentionned here: and also recognized that other enterprises start with BO right now
Maybe it is so because BO 4.0 is planned to be a new area (redesign) instead of fullfilling old roadmaps?
I am really interrested wether the OLAP paradigma will succed - I remeber that once it was the chalanger of BO’s ROLAP world and SAP did a crucial adaption to fit it with BW?
…just a thrilling subject for the next X years …waiting for 'qlickview to be a part of SAP '…?
BOByK (BOB member since 2012-01-15)
cough 6.1.a cough
6.5 was selected beta, 6.5.1 was the first general release.
If SAP get Hana as right as they say, then Xcelsius could be dealt a whole new lease of life and evolve into something different. The main concern with in memory solutions is what always existed with the good old RAM-based home computers of the 80s - how long to recover from a power spike? For a data warehouse, just power cycle the server and check it’s ok.
[quote=“Mark P”]
Not sure about Microstrategy but with Qliview recovery is very quick. The application loaded into memory resides on disk too. You will just need to reload it. 10 seconds?
twebber (BOB member since 2009-03-23)
Not alot of formal training in MS tools. I’ve spent probably three/four years with the MS stack of BI tools.
With Cognos I’ve had the most training and least amount of actual use. I developed lots of Cognos cubes and know their meta data engine pretty well. 3 year with Cognos but most of the project work was Oracle PL/SQL.
With BO I have had a two day course in universe design / build. One year hands on experience.
With Qlikview I have about two years of experience. One course on their Server Side management.
The rest of my BI training is dimensional modelling and some database stuff. All that doesn’t really add up to my 20 years experience in the industry. Hmm.
twebber (BOB member since 2009-03-23)
I’ve never seen or used Hana or other BO Olap. I’d like to see it.
twebber (BOB member since 2009-03-23)
Will or is Business Explorer be included in the latest 4.0 (Aurora) relase of Business Objects.
Also, is there a chance that Desktop Intelligence can return to the relase as it is the best desktop reporting tool ?
thanks in advance
ColumbiaIT (BOB member since 2011-06-15)
No chance, they are not going to put it back in after taking it out 
Nick Daniels (BOB member since 2002-08-15)
And SAP BusinessObjects Explorer is licensed, sold and installed separately.
Andreas 
(BOB member since 2002-06-20)
Explorer is pretty slick and I like it. Glad its still around.
As for DESKI; shame that SAP promoted a tool that goes against SAP Database. As we all know, NO ONE is going to use Universes for Crystal.
Not sure if Explorer will be a good case for me to keep BO. If not I guess Cognos or Q(as its being referred to) which are being used currently along with BO will earn another notch up as it continues to surpass BO. By end of next year I hope to still be doing BO, if not one of them will be the tool of choice 
Good Luck Aurora
ColumbiaIT (BOB member since 2011-06-15)
Explorer 4.0 was available to me under the Edge Platform. I was able to install and run it successfully. Didn’t get prompted for a license key either.
allovertheplace 
(BOB member since 2011-09-12)
This thread is an interesting read and I see there are people with various, differing requirements for their BI software, which can and can’t live with the move to 4.0 without DeskI.
For us we are definitely in the “can’t live with the move to 4.0 without DeskI” column. Yes we knew BO/SAP would abandon DeskI (however not for 10 years, at least up to 6.5 we hadn’t heard about that, only when XI R2 came out).
But to be honest we were relying on the DeskI feature set being included in WebI at some point. We do not see that and to us it seems WebI cannot do even the most basic formatting techniques that we are using in DeskI.
For example apparently you can’t even make a break in a table and then delete the useless column, ie have products and product groups, then break on product groups, then move the product group name below the product names and then delete the useless column. I mean if even that doesn’t work how am I expected to make any reports that can even remotely resemble what was possible in DeskI.
We are now working with Reporting Services and QlikView, both tools in my opinion are very poor compared to what BO DeskI was and could do, but apparently there isn’t any better. QlikView is good for analysis and agreed the front end is sexy, the back end scripts are tiresome and error prone to say the least, at the front end level it’s rather powerful, but also the syntax will drive you crazy (for a single set analysis formula, that is a bit more complex you need dozens of parenthesis and stuff looks like this sum(<Month={$($vMonth)}>Sales) even for simple formulas).
Reporting Services is more or less hardcore programming without WYSIWYG (although there are some interesting possibilities on the front end like conditional hiding of rows).
As we haven’t extensively reviewed all available solutions my question to the forum: does anyone know another reporting software that can combine data from different reports at the report level? For us this is the single most important requirement and we’re shocked again and again that apparently no other software can do that, they all require you to have some cubes or stuff in the background that already joins the data at the DB/DWH level - we do not want that, we want the report builder to decide what data should get into the report and what data needs to be joined in which way.
With RS it can’t do it (it only has a very simple lookup functionality in 2008 R2 that only works on a single column, but cannot join data over multiple dimensions).
With QlikView I gotta write some huge script for each “report” anyway whcih takes about 1000x more time than working with a BO universe.
We also looked at Microstrategy and it also seemed they cannot join data at the report level, it seems to combine data already in a single SQL query automatically voodoo style and in the few tests we did, we didn’t get the data in the way we wanted and also the single query Microstrategy created was much much slower than multiple queries in BO that then get joined at the report level.
Cognos might be able to do it, I don’t know, but it is financially not an option for us.
For us this is really a critical feature. I gotta admit we do not have any DWH knowhow here, so maybe we wouldn’t have this problem if we somehow built a DWH. For the foreseeable future however we don’t have it and we rely on multiple queries to retrieve data from different “modules” in our database (we mainly use BO for our customers as a reporting addon for our inhouse developed software that is mainly based on an MS SQL relational database)… and data from these multiple, different queries needs to be combined at the report level.
So… is there really not anything else out there that can join data from multiple data providers at the report level?
gerald 
(BOB member since 2006-02-14)