Impressions of Information Design Tool

It’s been well over two years since I got my first look at the SAP BusinessObjects Information Design Tool (IDT). However, that was pre-Beta, and the tool has undergone some changes since then. The version that went GA last September, 2011, wasn’t the same version that I originally saw in early 2010. Since the GA release, I’ve spent quite a bit of time with IDT, and would like to record my current impressions.

It’s important to keep in mind that Universe Design Tool (Formerly known as Designer), is a very mature product, and quite stable. IDT, on the other hand, is a relatively young tool, and, as such, has some quirks and bugs. I’ll address some of those here.

Let’s Start With The Good News

If you’ve read my earlier post on What’s New in Universe Design, you would know that the universe has a new structure, consisting of three different layers:

  • Connection Layer (.cnx)
  • Data Foundation Layer (.dfx)
  • Business Layer (.blx)

A Data Foundation can consume one or more connections. So you can bring tables from multiple databases into a single Data Foundation. The Data Foundations contains the tables, joins, and contexts. Optionally, you can build multiple Business Layers on a Data Foundation. The following diagram summarizes the new universe structure.

This structure adds some significant benefits over the old method of building universes. The ability to build a universe with tables from multiple databases has been a long time coming. And, the ability to build the Data Foundation once, and reuse it for multiple universes, means less work in the long run.

These layers are built in a local, or shared, project. One of the benefits of using shared projects is that multiple people can work on the same project at the same time. For example, one person can be working on the Data Foundation, while another is working on a Business Layer.

One of the new features that I really like is the ability to add calculated columns to tables. You can add a column to a table in the Data Foundation, based on a calculation using columns from that table, or constant values. Once created, the column appears as just another column in the table.

You can also create time hierarchy columns automatically. This is so much better than the old Automatic Time Hierarchy feature.

I also like the new Families feature. A Family is a collection of related tables. Families can be identified by colored backgrounds and fonts. You can create as many families as you want, and add as many tables as you want to each family. A table can belong to one, and only one Family. You can then search for tables by Family, Name, Columns or Contexts. You can search by wildcards as well.

Another feature that has good potential is that Lists of Values, as well as Prompts, can be built as standalone objects in the Data Foundation. This means that they can be reused in multiple objects in a universe. But they can also be reused in multiple universes. However, the process of creating them is cumbersome and quirky. The various options aren’t well organized, and require too much experimentation to get them right. But once built, they work quite nicely.

In the Universe Design Tool (UDT), we’ve always had an issue with testing as we build. We could test in Desktop Intelligence, or, more recently, in Web Intelligence Rich Client. There is a Query Panel in UDT, but it’s mostly useless. However, in IDT, we can fully test as we go, without relying on another tool. We can build, and execute, queries, within the Business Layer. These queries can be saved in the universe so they can be used again. Once we run these queries, we can analyze the results with sorts, filters, and changing the order of the columns.

A feature that completely eliminates the need for Linked Universes is Business Layer Views. When you create a Business Layer, that becomes the Master layer. You can then create Views of that Business Layer, choosing which folders and objects to expose in that view. You can then grant rights to those views for various user groups.

Not All News is Good News

As I mentioned, IDT is a fairly young tool, and, as such, has some bugs and missing features. In fact, some of these missing features can cause huge problems.

I always recommend applying formatting to every numeric object in a universe. After all, we don’t want to force the users to apply formatting every time they use these objects. In UDT, we could select multiple numeric objects at once, and format them all at the same time. However, in IDT, we have to format objects one at a time. And the process of creating a format is very weird. For example, here is what the format for typical currency might look like:

That’s right. You can’t simply type in the format. You have to use the objects on the left, and add the items to your format. And, to make matters worse, once you create a custom format, you can’t save it. If you want to format another object the same way, you have to create the same custom format again. It’s very time consuming, and will likely discourage many people from formatting objects in the universe.

In the UDT, when we are using the @Select or @Where functions, we can check a box to “Show object SQL” to see the full SQL of the object. It’s very helpful for troubleshooting. Unfortunately, this option is missing from IDT. There is no way to see the full SQL when using these functions.

Another missing feature that is, in my opinion, a serious issue, is the ability to change a dimension to an attribute (Formerly known as detail), or vice versa. For example, if you create an object as an attribute, and later decide that it needs to be converted to a dimension, so it can be included in a Navigation Path (formerly known as Hierarchy), you can’t do it. You have to recreate it as a dimension. This, of course, would cause existing reports, that use that object, to fail. In my opinion, this is a huge problem.

The process of converting UNV universes to the new UNX format is very buggy. I have seen parameters get changed, and universe settings also get changed. As my friend, Dallas Marks, discovered, even the eFashion universe cannot be converted.

Now, to be fair, SAP seems to be doing a good job of listening to our concerns with the new tool. I have submitted quite a few ideas on the SAP Idea Place, and several of them have already been accepted. That impresses me. Now we’ll see how quickly these issues are addressed. In the meantime, I suggest keeping your current UNV universes in their current format, and creating new universes in IDT.

If you’ve worked with IDT, what is your experience so far?

Advertisements

37 Responses to Impressions of Information Design Tool

  1. jerimiahbaldwin says:

    As you said, the inability to convert dimensions to attributes (and vice versa) is a real killer. I am curious of the differences in behavior between XI 3.1 dimensions/details vs 4.0 dimensions/attributes. While I’ve not explored this in depth in 4.0, I’ve used them enough in 3.1 to see certain pros/cons to classifying an object as either a dimension or detail/attribute. For example, as you said, one might want to change a detail/attribute to a dimension for the sake of including in a navigation path = detail/attribute objects cannot be used in navigation paths. Also, in XI 3.1, the table block rows are grouped by unique *dimension* value, not detail/attribute. Therefore, if you have a table block with both dimension and attribute fields, and a given dimension value has multiple distinct attribute values, the table block displays a #MULTIVALUE error. Is this functionality the same in 4.0?

    Previously, I’ve casually leaned toward classifying objects as dimensions to lessen this affect for users. Today (in 4.0), this decision carries a lot more weight. I’d be happy to see that this be changed someday.

    • The limitations on details in 3.1, is the same in 4.0. The cannot be used in a Navigation path, and they cannot be used to merge data providers. However, details/attributes can be added to tables in Web Intelligence, without limitation. I wrote about that here, and it still applies.

  2. Thanks for posting this, Michael. I’m anxiously awaiting access to Feature Pack 3 to see how many of our mutual concerns are addressed.

    • I agree, Dallas, but I’m concerned, after reading some of the early experiences of those installing the ramp up version of FP3. I’m hoping the GA version is more stable.

  3. K von Murphy says:

    Hi,
    Thanks for the report. I’ll wait a few service packs before making a real universe with IDT. The old Designer may be the only tool in BI 4.0 that is not full of bugs, so I keep it for the moment, I don’t really need the new features so much :-\

    Anyway, I played with it and I was shocked by the loss of drawing speed: a (huge) universe needed many seconds to appear on the screen in IDT, although it was immediate in Designer.

  4. KhoushikTTT says:

    Hi,
    any insight of how IDT works for OLAP universe. Im quite interested with the multiple data connections. Could this option be used for a OLAP based universe.

  5. Aina says:

    Did you manage to find Multiple SQL statements for each context option? Is it enabled/disabled by default?

  6. First of all, kudos to you Michael for your blog.

    But now to the IDT: To me the IDT is cumbersome, workflows are not intuitive.
    – Why oh why is it called a “project”, but when published it becomes a UNX? Does not make sense.
    – Why do I have to create a shortcut for a connection anyway instead of directly using a secured connection?
    – And why can’t I utilize 3-tier mode (it would be very useful to not have your workstation setup for each single connection locally).
    – Creating a simple universe takes at least double as much time now : -(
    – Menues are hidden, stacked, etc. the whole GUI is just a maze, even worse than Webi 4.0 has now become.

    SAP really needs to get a really good GUI designer on board to straighten this mess out.

    • Andreas, I completely agree. One of the ideas for this tool was to reduce the amount time needed to build a universe, but that idea failed.

    • Dallas Marks says:

      Andreas,

      I think some of the UI is a result of Information Design Tool’s Eclipse underpinnings. Having survived the transition to Microsoft Office 2007, I’m OK with learning new ways of doing things, but I agree that some key workflows and features of the classic Universe Design Tool are currently missing. I’m not seeing them in Feature Pack 3, so I would hope that there’s time to get at least some of them into Support Pack 4 rather than just bug fixes.

      The BusinessObjects guru Yoda said that “you must unlearn what you have learned”. I imagine new designers learning the IDT for the first time will have an easier time than those of us unlearning the Universe Design Tool.

  7. Pingback: Object Formatting with the Information Design Tool

  8. Pingback: BO 4.x Migration Challenges « My life as a BI consultant

  9. Brian Wolf says:

    I’m finding IDT suffers from the same thing WebI does: too much flash, not enough substance. While you can shut off the animations in IDT to improve performance, there’s still performance issues with dealing with tables which have large numbers of columns. Takes the screen way too long to repaint. Waiting 10 to 15 seconds between clicks isn’t terribly productive.

    Here are some props: Experienced the “layered” approach to meta data design in OBI 10.7 (yup, years ago…). I found it to be redundant work. Fortunately, the work in IDT is compartimentalized, not redundant.

    • Gary says:

      I just started playing with IDT and I had the same issue on performance for tables with large number of columns. Do you have any suggestions on how to resolve this? I can’t get any work done.

      • Gary, sorry for the slow response. What kind of performance issues are you experiencing? Have you resolved it?

      • Gary says:

        The performance issue I’m experiencing is occasionally when I open the data foundation or business layer in IDT, the pane keeps on refreshing by itself nonstop. I have to let it refresh for about 5-10 minutes OR kill the IDT process and start over. That’s quite irritating.

      • That’s interesting, but not something I’ve seen. Try opening the same universe in IDT, on a different machine, and see if it still happens. That way you can narrow it down to the universe, or the machine.

      • Gary says:

        Thank you for your suggestion. I’m planning to put Windows 7 on my laptop and then I’ll see if it improves.

      • Nick says:

        I had performance issues with the unx. My unv would process a query in under 2 seconds and the unx took 5 minutes. In the end we discovered a permission setting that caused it. Too funny. I keep asking myself why it would affect it.

  10. zulfikar says:

    My experience using IDT was not that pleasant,like most of them above.while working on universe found it was really difficult to create contexts in the universe.Finally i had to drop idea go with alternate solution.Then there are lot of steps before publishing it to the Repository which really cumbersome.For example Syncronization of the Unx file with repository file.Then while importing the universe you have to create new project which really tiring.According to my opinion SAP is looking its friendlyness as it gained in earlier version.

  11. Jansi D says:

    Ah, they have to fix at least the major issues before they make the product go live. In other words, Beta testing should be effective/beta testing feedback should be considered.

  12. Raoul Labega says:

    Michael,

    I have been experimenting with multiple business layers on top of 1 data foundation and I cannot see the advantage of this feature. When I create multiple business layers on 1 data foundation and publish them, they all have their own local copy of the 1 data foundation. So in fact I have multiple data foundation now. If I then change one of these copies, I have to manually change all the other copies. I could not find any options to change the data foundation on an existing business layer.

    So using shared project mode with multiple business layer designers is nice as long as the project is not publish. Because after it is published there are multiple copies of the data foundation.

    Maybe I am missing something here, hopefully there will be some changes in coming releases. But this option to create multiple business layers (and multiple copies of the data foundation for free) is not something I should recommend using.

    • Raoul, That’s another thing that, in my opinion, wasn’t thought through very well. I believe the idea was to use that “feature” for versioning. But it causes more problems then it solves.

  13. Emilien G. says:

    Michael,

    Thank you very much for your feedback, which is really interesting and gave weight against SAP standard communication.

    I am actually studying the opportunities to work with IDT (for the multi sources aspect) rather than integrating data directly in SAP BW before using the BICS connexion.

    In my opinion IDT is not really powerfull to raise the risk of a bad quality of data. What do you thing on this subject ? What if data from SAP, from BW and from an excel serveur arrived with a poor master data quality ?
    As an exemple if a date is not formatting in the same way in every databases ?
    Is there any tool within the IDT to deal with this kind of issue ?

    Regards

  14. Gary says:

    HI Michael,

    1) In IDT, we are dealing with projects. Is the purpose of the Project for each universe? Or we can create multiple uinverses within a project?

    2) The new concept of Project Synchronziation allows multiple Universe Designers to work on a project, which is good I think. But I think there is still no way to prevent multiple Designers to publisher universes to the repository, right? I can publish my copy and my colleague can publish his copy 5 minutes later to overwrite mine?

    3) What is the purpose of Date and DateTime data type for each object? I can’t see any difference?

  15. Ton Vermeij says:

    I wonder if the usage of IDT has improved in 4.1?

    How is the performance for developers? And have little ‘things’ as mentionned above like changing dimension objects to detail objects not possible? Also for changing the format of objects groupwise.

    • Most of the missing features that I complained about in 4.0 have been added in 4.1, making a huge improvement. I am generally pleased with IDT in 4.1. The only area where I am not happy is the area of contexts. Contexts were supposed to be simpler in IDT, but have become significantly more complex. In Designer, every join has one of two relationships with a context. It’s either a member, or not. But in IDT, a join has one of 3 relationships with a context: Included, Excluded, Neutral. I have found that this makes contexts much harder to work with. So, when I have a large, complex universe to build, I do it in UDT.

  16. Andrew Dale says:

    in UDT and IDT in BI4.0 and BI4.1, is there a limit to the amount of text put in the Comments field of the universe Parameters box?

  17. vivek pandey says:

    Hi Michael,

    Could you please explain benefits of Edit table column properties in IDT, such as a option to change data type of a column in IDT, since we have option to cast and object definition in UDT as well.
    In addition, how IDT process this request?

    • Hi, Vivek;

      That’s an excellent question. I can’t think of any reason to change the data type of a column. The data type is read from the database, and changing it in IDT isn’t going to change it in the database. So I don’t see any purpose in that. However, I can see some benefit in hiding columns in the Data Foundation. If you have a table with 100 columns, but you only need 10 of them, it can be nice to hide the ones you don’t need, so the table appears smaller.

  18. Pingback: Review of Information Design Tool 4.1 | Michael’s BI Safari

  19. Pingback: Object Formatting with the Information Design Tool | Dallas Marks on Business Intelligence

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: