This Essbase discussion board is provided as a free service and dedicated to all the Essbase professionals out there!
  << Previous Topic | Next Topic >>Return to Index  

Question for SV trainers

June 2 2017 at 9:24 AM
No score for this post
Tim Faitsch 
from IP address 166.76.0.1

How do you explain the whole "member name only" option actually displaying aliases because it's set someplace else issue? I've tried explaining it several ways but each time I feel more frustrated. I get blank stares in class even after I demonstrate how it works.
Thanks,
Tim

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
AuthorReply

Jeff McAhren

165.225.34.125

Sorry, no easy solution

No score for this post
June 2 2017, 10:41 AM 

It *is* confusing, and there is no good way to explain it to users. Sorry.

I just point out the two locations that control this, and the users have to fight through it. Too make it as difficult as possible, the Change Alias button is on the Essbase ribbon, and the Options button is on the Smart View ribbon.

Be glad you only have Essbase, because when using Planning connection types, it's different. And for HFM connection types, it's different there, too. Users call me all the time, confused about how do do XYZ, because they know they did it before (but they can't do that in HFM, or using a Planning connection, or ..).


    
This message has been edited by jmcahren from IP address 165.225.34.125 on Jun 2, 2017 10:44 AM


 
Scoring disabled. You must be logged in to score posts.Respond to this message   

Jeff McAhren

165.225.34.125

Power Users

No score for this post
June 2 2017, 11:00 AM 

BTW, everywhere I've worked with the Add-In during the first 15 years of my Essbase career, power users came out of the woodwork, which is awesome.

I've worked with Smart View for the last 8 years (at my last three organizations), and have had a really hard time developing power users. I've been at this organization for five years, and I have three power users.. me, and two guys that sit in financial reporting (the application owners). That's it.

Everyone else struggles with the tool. Most can't even create adhocs from scratch, they can only refresh existing templates, and I don't think it's because they aren't smart people, or aren't receiving adequate training. I think it's because the tool is difficult to use.

So you have that to look forward to.

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Jared

149.128.8.245

Struggles

No score for this post
June 2 2017, 12:25 PM 

Just curious, what are they struggling with?

We've just completed a company-wide move from the Classic Add-in (which had been in use here since 1999) to Smart View 11.1.2.5.620 (Essbase only) and it's gone pretty smoothly. We had an initial training session for manager-designated SMEs and training for everyone. The SMEs are primarily responsible for doing report conversion and being "champions" of the product. So far, so good.

 
Scoring disabled. You must be logged in to score posts.Respond to this message   

Jeff McAhren

165.225.34.125

Re: Struggles

No score for this post
June 2 2017, 3:25 PM 

I think it comes down to a lot of little things. A couple of examples:

On the subject of aliases, the Planning connection type cannot do "member name and alias", users must pick one or the other. The HFM connection type *can* do "member name and alias, but it concatenates them together in one cell with a dash in between, and it does it for every single member on the ad-hoc (not just the row dimension(s)). Essbase, as we know, givees a column for the member names and another for the aliases, but for only the row dimension(s). None of this is intuitive configure (as documented in this thread). You can type member names directly on an adhoc grid, and Essbase will recognize it, and change it to the alias when refreshing (i.e., type P01, and after refreshing, it is "Jan"). HFM can not, it throws an error. When every connection type functions different, it can be frustrating for users.

Another one is the POV toggle button that hides/unhides row 1, but only for the Essbase connection type. We instruct users not to use the POV bar, but they do anyway, and they inadvertently delete members from the page when deleting columns, or they duplicate members when copying columns, which results in errors that they cannot troubleshoot themselves. The other connection types have the POV toggle too, but they don't hide row 1, or place the POV members on the ad-hoc worksheet. I know this sounds minor, but it's the 100 little things like it that make Smart View difficult to use.

BTW, what do you think about the drop down buttons on POV members.. like or dislike?

We've also fought some bugs. We add a few hundred new construction projects to the entity dimension each month. We trained users that if they have projects on the rows in their ad-hoc, they need to zoom out on to the division or office level, and zoom back in to catch the new projects that have been added since last month. But when zooming back in, a Smart View bug caused it to add (x) rows between each member, where x equals he number of rows in the last grid. After zooming, users only saw one entity, and didn't realize they had 200, but they had 175 blank rows between each one. So they had to create a new ad-hoc each time, or clear Smart View metadata.

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Pete

124.188.5.16

Re: Struggles

No score for this post
June 2 2017, 7:29 PM 

I agree completely with Jeff - if you've got a single product you're kinda fine, but if you're trying to train across the entire essbase\planning\HFM stack the (meaningless!) inconsistencies are a pain in the arse.

In many ways it seems like the oracle development guys are afraid of standardising to one particular method (imagine the outcry on this board if they went to the HFM formatting!) but...it's to the detriment of the entire product.

And the POV buttons are a travesty. The default is awful, the behaviour inconsistent, the drag and drop output is a nightmare (why would you include the top level dimension name as well in a list of 1!)

All that being said? I like Smartview. I really do. I just occasionally need to remind myself.



 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Tim Faitsch

166.76.0.1

Re: Struggles

No score for this post
June 2 2017, 3:48 PM 

This is the opposite of our experience. I'm curious how long was the conversion and how many users were you converting?

It seems our greatest struggle has been with large retrievals. Many of our users essentially want to create their own hierarchies on the fly. They do this by pulling large retrievals (think 20k+ rows by 200 columns). Then they use filters to add up various combinations of members. I had no idea so many people were doing this until their 2 to 5 minute long pulls turned into 20+ minutes in Smart View -- that's when they didn't crash.
Tim

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
TimG

99.66.117.90

Re: Struggles

No score for this post
June 3 2017, 8:50 AM 

I see problems with that too, Tim.

That said, users performing "data dump" retrieve almost always points to an opportunity to enhance the application, whether that means formulae, alternate hierarchies, additional dimensions or attributes. I doubt this is teaching you much. :)

It might sound perverse, but I actually rather like that the performance problem calls these to my attention. I suspect users would go on quietly using their workarounds for years otherwise, unaware that the cube could do it for them...

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Tim Faitsch

166.76.0.1

Re: Struggles

No score for this post
June 5 2017, 9:52 AM 

I try talking them out of doing it or talking about redesign but they answer back, "it worked in the old version, telling me to change is not an acceptable answer". How am I supposed to respond to that? In their minds I'm taking essential functionality away from them. Functionality they've had for over a decade.

Tim

 
Scoring disabled. You must be logged in to score posts.Respond to this message   

Jeff McAhren

165.225.34.125

Re: Struggles

No score for this post
June 5 2017, 10:46 AM 

Yep, you can't respond to that.

You're probably already aware, but you might be able to improve performance a bit by disabling undo, not using styles, unchecking the 'move formatting on operations' option, and unchecking the 'preserve' member options.


 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Tim Faitsch

166.76.0.1

Re: Struggles

No score for this post
June 5 2017, 10:52 AM 

Yep, we did all that. The users with large pulls also have no choice but to turn on "manual calculation". Seriously. This product makes power users turn off one of the most basic of all functions in Excel. It's shameful IMO.

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
TimG

99.66.117.90

Re: Struggles

No score for this post
June 5 2017, 4:23 PM 

I hear you, but not even if you tell them "we will provide a *better* solution"?

Say they are pulling all the items that meet a specific criteria. Just for example, "Organic" or "Made in the USA". You systematize that with either a dimension or an alternate rollup and now a) users never have to wonder "is my list up to date?", b) it's really fast, c) everyone who tries to query the same thing agrees on the numbers and d) you don't have to recreate your non-additive formulae (Gross Profit %, or whatever) with Excel math and hope you got it the same as the cube.

Again, I doubt this is telling you anything, Tim, and I'm *definitely* not defending Smart View. By killing the Classic Add-In Oracle created a market for the (superb) Dodeca Excel Add-In. But even the Clasic Add-In wasn't built for data dump retrieves (row and column limits). Essbase - in general - is crap at acting as a bulk data source.

As long as you find an alternative way of meeting the requirement, no reasonable user can blame *you* for what Oracle choose to do, surely?



 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Tim Faitsch

166.76.0.1

here's what they do...

No score for this post
June 5 2017, 5:01 PM 

Someone creates some monstrous retrieval. It's got months, quarters, several scenarios, tons of divisions, gobs of products and a handful of accounts. They have another sheet with lookup tables. Then a couple dozen people sit in a room, everyone with their laptops and the sheet. Someone says, "pick x, y, z division in your filter for x, y, z products " and then they talk about the numbers. Since the data pull is a one time deal and Excel filters/adds everything super fast, it's a very good solution for what they're doing.

So essentially they're creating a dynamic list of divisions and products to add up each time. And the lists change all the time -- even within the same meeting. They might talk about one list and then say, "now remove z". I'm guessing this is something that Dodeca could do.

What would be really cool is if the POV in Smart View was multi-select. So imagine being able to, on the fly, add up 15 divisions and 10 products and look at your P&L for that. I know that member on the fly is something Essbase development team has talked about and it's possible, but painfully slow in MDX...

Tim

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Pete

210.11.186.210

Re: here's what they do...

No score for this post
June 5 2017, 9:32 PM 

Hey Tim,

yeah - that one is hard.

it would be 'possible' to build some kind of VBA excel front end to use the smartview mdx output to build queries creating dynamic members.

Tim (G) wrote a very good article on the actual MDX.
http://www.cubecoder.com/use-mdx-for-custom-roll-ups-that-respect-ratios-and-variances/

But a possible solution and a good (and supportable!) solution are poles apart - and that is before you have to deal with the performance impacts.

Dodeca could do that for sure - get Jason Jones or Tim Tow talking about dynamically creating MDX queries based upon other queries and you're in for a fun few hours. That said, it's another business change.

P




 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Tim Tow

68.35.22.97

Re: here's what they do...

No score for this post
June 18 2017, 11:01 PM 


Pete,

Are you calling Jason and me out as 'geeks'? If so, we worked *hard* to earn those badges! <BigGrin>

You are correct; Dodeca can do that. After all, that is one of our taglines - Dodeca Does It.

We just did a couple of webcasts in the last few months that covered MDX in Dodeca including one webcast where we showed a multi-select point-of-view selector used as input for a 'member on the fly' calculation. Here are a couple of links:

Introduction to MDX for Advanced Reporting in Dodeca - https://attendee.gotowebinar.com/register/8796395769046177283

On the Fly Member Calculations with MDX in Dodeca - https://attendee.gotowebinar.com/register/7236222854909240835

Tim

Tim Tow
Oracle ACE Alumni
Applied OLAP, Inc

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Tim Faitsch

166.76.0.1

MDX performance

No score for this post
June 19 2017, 8:59 AM 

How has performance been with the MDX member on the fly? I added a simple A + B member and it turned a 30 second report into a 20 minute report. Independently they are fast but with the addition, it was painfully slow.
Tim

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Tim Tow

50.247.46.34

Re: MDX Performance

No score for this post
June 19 2017, 6:09 PM 


Hi Tim,

Like so much else in Essbase, there are so many variables to consider that performance on one cube and query can't really be compared to the another. How many cells were you trying to return that had to be dynamically calculated? Would it be possible to slice down the amount of data so a user could focus on what matters to them?

Tim

Tim Tow
Oracle ACE Alumni
Applied OLAP, Inc

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Tim Faitsch

166.76.0.1

Re: MDX Performance

No score for this post
June 20 2017, 9:44 AM 

I was just asking in general if you had seen issues. My specific issue was with a large MDX extract that ran fine when pulling two members independently, but when I added them together took forever to run. I think I'd be even more concerned with doing it on multiple dimensions in the same query. So adding up a group of stores and a group of divisions and a group of products -- which is basically what my users want.

We had initially tried using MDX to handle this in our custom built java based front end but once you get it back, you're stuck using MDX for everything and can't go back to using a grid for zoom-ins and such. Would have been a huge redesign effort for us.

I just had a flashback to about 1998 when I was at a client who had created a multi-select in VBA/Excel and wanted that recreated on the web. I was able to successfully (but painfully slowly) recreate this by using hidden frames and doing individual queries on the multi-selected members then using javascript matrices to add everything up and present the totals to the user. Feels like a lifetime ago and yet the requirements haven't changed one bit...

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
Tim Faitsch

166.76.0.1

actual results

No score for this post
June 20 2017, 11:03 AM 

Output of the MDX is 462,699 rows.

I have four members as column headers on the first pull:
MaxL DML Execution Elapsed Time : [91.8] seconds

I add two of the four members as one dynamic member and pull the remaining two members in the column header on the second pull:
MaxL DML Execution Elapsed Time : [1142.37] seconds

It looks like it takes a very long time to add those two columns together. Would be interesting to know how this works. Might be whitepaper material -- where it breaks, how it's working behind the scenes, etc. Maybe someone has already written one?

Tim


 
Scoring disabled. You must be logged in to score posts.Respond to this message   
DrFosterman

159.53.174.143

Leaves?

No score for this post
June 20 2017, 11:27 AM 

By any chance are you using leaves? if yes, does the log specify that it is expanding one of the dimensions in the second query but not he first query? Just wondering.

 
Scoring disabled. You must be logged in to score posts.Respond to this message   
 
< Previous Page 1 2 3 Next >
  << Previous Topic | Next Topic >>Return to Index  

RSS feed for this forum - http://www.network54.com/Forum/58296?xml=rss. Please email hypess (at) gmail.com, if you have any questions/feedback/issues.