Skip to Content

BI 4 BW and ECC: Where are we really? (1st: Explorer)

Disclaimer: These are the results of some of our tests, the fact that they’ve worked for us does not guarantee they are officially supported by SAP.  Also, I have purposefully excluded solutions involving HANA or BWA as I mostly wanted to address SME needs, and most of my SME customers are not planning on implementing a HANA or BWA scenario within the next year.  Also, there are enough HANA-related posts already, I have no added value adressing these issues.

This is the first of a series of posts in which I will try to go over where we are today in a context of SAP BI used in conjunction with a SAP BW and ECC platform WITHOUT SAP HANA.  We see a lot of posts about what is coming, we see a lot of promises and we see new SP announcements every month or so, but what is going right (and wrong) with SAP’s BI platform today?  Particular emphasis on BW integration.

To give you a feel for the breadth I will try to give to my posts, my team and I have been involved with a dozen different BI4 implementations on BW starting from the original ramp-up in 2011 (GA in September 2011) up to a few SP6 installs .  We are also involved on the 4.1 ramp-up, but I think I’m not supposed to discuss that here (I’ll have to read-up to make sure).  All these were centered on a BW platform although we did use IDT (new universes) for some requirements and we have been through extensive developments using Webi, Crystal Enterprise, Analysis for OLAP and Office (1.2 to 1.4), Explorer, iOS mobile apps, Design Studio and Dashboards.

As for yours truly, I have been in the BI field for the past 15 years, starting with Crystal and Seagate Info (for people not tracking names changes, Seagate Info became Crystal Enterprise which then became BO XI and finally SAP BI4) and also intensively working with BW for the past 9 years (BW 3.5 –> 7.31).

So what do I plan to show?  I plan to go through screenshots of the different tools showing what really improved user and developer experience when compared to BEX, but also what doesn’t work as well and some lessons learned.  I will only cover GA versions (mostly SP6).


OK, enough talking, let’s start.  The first tool I plan to cover is Explorer 4.0 SP6.  As you probably know, Explorer is not officially supported on BW. 


We only have two available sources of data, namely Excel files and .UNX universes defined in IDT (“classic” universes, .UNV, are not supported but are part of the latest roadmap I saw update: see note 1858836 – How to enable .unv support for Explorer in BI4.0 SP06).  /wp-content/uploads/2013/06/sources_228305.png

So when you only have two options, you go along and try the one that seems the most practical (not Excel).  So lets build a .UNX on top of a BW cube using IDT.  IDT_program.png

Create your Universe:

I won’t go through the detailed steps, but quickly:

1. I used the SAP BW JCo connector for a connection on standard cube 0FIGL_C10

2. Created a data foundation on the preceding connection, star schema is automatically generated as seen below:


3. Created a Business layer to define what I want to see in Explorer, the business layer pretty much recreates our cube dimensions automatically.  The main thing we’re missing are hierarchies:


4. Finally you publish your universe to the BI platform repository.

Total, these first few steps only take about 5 minutes as everything is pretty much automated.

Create Information Space

Now comes the fun part, publishing the result in Explorer.

First you open Explorer an go and select the Universe we just created:


The next step is to give it a name and some basic properties, but the interesting part comes in the second tab where you select the characteristics you want:


You notice the third tab?  Scheduling.  Very basic, but you can refresh your information space automatically every morning through here. 

Finally you save and go have a look.  This is what I got:


You’ll notice that some characteristics have texts, others not…  For this I cheated a bit since there is no “show key/text” option in Explorer, what I did is slightly modify the SQL in the Business Layer of the Universe for the specific keys.  This SQL modification is quite simple and can be done through a basic drag and drop of the text field and adding a “+’ ‘+” in between the two values to generate a space.

So here we are, Explorer on BW. 

Some drawbacks:

1. hierarchies are not supported

2. No exception agregation

3. Based on Infoproviders only, not Bex queries (although I haven’t tested if a BEX query as an Infoprovider in 7.3 could work, will get back to you)

4. No BW-based authorizations on data as the Explorer index is generated once, no back and forth with the BW server at runtime.

Update:  More extensive information on what is supported or not can be found through this blog from Ingo:–latest-documents

Are these show stoppers? I think not, but if one of these is very important, clearly Explorer is not for you.

Something else to keep in mind, this solution actually works out in the context of Lumira (ex-Visual Intelligence) and Predictive Analytics as well.  These two solutions do not support BW either, but they support .UNX universes.  Update: as indicated in comments by Ingo below, this is not supported.  But last time I tried it worked…  No guarantee here.

Now the one thing that would really make me happy? Integrate these solutions directly on BICS.  BICS can deal with hierarchies as separate levels (will discuss this in my Webi piece), so aside from Exception aggregations that probably won’t work in this model, I don’t see what the show stopper is.  Everything else on the BI platform can now deal through BICS.


So what did you think?  Do not hesitate to provide comments below.  If you have a specific BI tool you’d want me to go through first do say so.  Right now I’m planning on going through Webi next, but I’m open to suggestions.

Hope to share experiences with you all soon.

You must be Logged on to comment or reply to a post.
  • Hello Eric,

    I would like to correct some statements of your blog here:

    – you mentioned that Explorer has only two available sources : Excel Spreadsheets and Universes (UNX), but you forgot to mention BWA and HANA (BW on HANA), which are important especially in a BW landscape. Explorer is able to directly leverage the data from BWA and HANA as well which also allows Explorer to go far beyond the data volume that is possible with a Universe.

    – you listing the limitations of the relational Universe but you might want to list elements that are not available in the InfoProvider as well, such as restricted and calculated key figures and custom structures.

    – You then suggested to have Explorer leveraging the same interface – BICS – as the other BI client, but that does not make much sense. Why ? BICS is a request based interface, means you navigate in the data, you asking for a subset, you get the subset of data. Explorer is a discovery tool where you want to use millions of rows – so a request based interface would not make much sense here.

    – You also mentioned that Visual Intelligence and Predictive support UNX – correct they do support UNX, but the support for BW as data source including the support for BW via UNX –  is explicitly declined.

    – For Visual Intelligence and Predictive and Explorer customers should look at the option to leverage BW on HANA and automatically create HANA models on their BW models and then leverage Visual Intelligence, Predictive or Explorer on the HANA model.


    Ingo Hilgefort, SAP

    • I have sussessfuly tested the option to link Explorer to a BWA index based on a CUBE and also based on a QUERY (Set as an infoprovider).

      I used the instruction found at this SDN document:

      I do have a question, when you Eric says:

      4. No BW-based authorizations on data – what does that mean?

      if we create an Universe using relational connection from BW (Jco) we loose the security model from BW?

      Thank you

      Rodrigo Silva

        • So, if I have a security model that is based on a Hierarchy Node , I will loose the hierarchy for reporting on the Universe but the access will still be checked? somehow?

          In your case, we have security roles by Cost Center hierarchy nodes and Cons Unit Nodes…

          thanks for you quick reply.

          Rodrigo Silva

        • Hi Ingo,

          I’ll reply to your previous comments later.  But technically you DO lose the BI authorizations. 

          1.  BI authorizations require a Bex variable to work properly and thats officially not available. 

          2. Afterward, if the tool actually checks the authorizations in BW it won’t filter appropriately (that’s done by BEX) and the user would have to figure out what he has access to…  until then he’d get an error message (very detailed, saying: “No authorizations”).

          Because of point 2, I don’t see how these can be used. 

      • Hi Rodrigo,

        I updated the text to clarify what I originally meant:

        4. No BW-based authorizations on data as the Explorer index is generated once, no back and forth with the BW server at runtime.

        Our discussion then went another way (.UNX authorization support) which I’ll dig into over the course of the next week.



        • Hi Eric,

          that statement is not correct.

          Explorer is using the BW authorizations. Correct there is no BW security AFTER the index has been created – assuming we are talking about the UNX based connectivity, but that is true for any other data connectivity as well assuming they are UNX based becasue Explorer creates the index and that is the nature of it.

          The answer to that is the usage of personalization in Explorer.

          Eric – based on the comments you receive here I would suggest you re-work major parts of the blog as right now there are a couple of statements in there that are technically misleading.


          Ingo Hilgefort, SAP

          • Hi Ingo,

            Thanks for all your comments, but I think I’ve addressed the items that were misleading.  My objective was never to create a How-to blog, but rather what I experienced using the tool. 



        • Thanks Eric, maybe my question did belong on this post, since I was not talking about Explorer anymore. I will test the security using the UNX from Cube, and I will reply.

          I have a security role based on hierarchy nodes and I want to see how that works, since I know I will lose the hierarchies for reporting on the Universe.

          My other option is to go back to BICS for my reports, but I do have an issue that after I choose the BW query to create a WEbi report, it takes 25 to 35 min to show the fields. I heard that this will be solved at BI 4.1, but I am losing my customers at the moment and need to delivery something before they give up BOBJ.

          • Hello Rodrigo,

            on the timing part when creating a report with Web Intelligence it looks to me that sizing might be a root cause here as we have seen that on other deployments as well.

            The important factors for the “design time’ are the number of elements in the BEx query but also the sizing of the involved components is important


            Ingo Hilgefort, SAP

          • Hi Ingo, after several weeks with the problem (Waiting 35 minutes to open webi after selecting the BW query) we found the cause and solved the problem.

            Debugging what was happening in BW, we found that the webi interface was reading all the hierarchy master data table for WBS (and we had 9 active hierarchies). that was causing the long wait time.

            To solve the problem, I deleted 4 obsolete hierarchies and now it loads in less than one minute.

    • The author is probably talking about non accelerated version of Explorer. Hierarchies are fine, because you can filter the way you like to get to the lowest level details or vice-versa.

      Working on UNV universes is fine, down the road we need to adapt to IDT and I do not see any disadvantages of IDT over UDT. Even it has more features.

      But what I feel is a big disappointment is distinct count of higher levels. For example if I have three levels of granularity, Explorer will present you with only the count of least level granularity. However the calculation engine calculates the distinct count, but you will only be able to see when you click the drop down button of the Facets when trying to plot data. Besides sum, count is also important thing business users like to see.

      Getting the SQL query. We had to cheat by purposely by writing little wrong SQL and Explorer generates error message immediately. However this requires publishing universe twice.

      • Hello Rajananda,

        Not sure why you stating that “hierarchies are fine”, as hierarchies are not supported when using the UNX or UNV with the relational connectivity.

        Hierarchies are only available as a filter in the combination with BWA.

        On your second part where you talk about the different level of aggregation – remember that when Explorer uses a Universe you are basically creating an index of the data, so calculations won’t happen on the fly anymore, the index has been created at that time.


        Ingo Hilgefort

        • When I said Hierarchies are fine, I meant you don’t need Hierarchies in Explorer to see various levels of data, because you have all the data in one interface you can filter the way you want.

          “On your second part where you talk about the different level of aggregation – remember that when Explorer uses a Universe you are basically creating an index of the data, so calculations won’t happen on the fly anymore, the index has been created at that time”

          I know and not expecting calculations on the fly. It’s pre-calculated in the universe, but Explorer won’t simply show. It’s a known issue and has been submitted in the Idea Place year ago. No solution so far…

    • Hi Ingo,  I think I adressed your issues in your first comment through some updates in the text.  Namely:

      1.  I was purposefully avoiding HANA and BWA solutions, I now clearly state that.

      2.  I was pointing towards the limitations that made us question this solution the most, I will add a link to your more detailed document you recently published.

      3.  You’re probably right.  But generating an index using the BICS connection instead of going through the whole .UNX process would feel more natural to BW users.  I was stating a wish here, but can’t go further as there are probably multiple technical limitations in the backend…

      4.  Added an update about Visual Intelligence and Predictive.  Found the reference you state in the documentation.  You’ll notice I never said it was supported though, only said it worked 😉 .

      5.  This brings us back to my newly instated disclaimer.



      • Hello Eric,

        on #1:  I always think that even if you want to “avoid” certain topics that at least they still should be mentioned as otherwise people will get the impression that it is not an option.

        on #3 : For Explorer the normal process is BWA and HANA based as a UNX or BICS based interface would not provide the necessary volume of information.


        Ingo Hilgefort

  • Hi Eric,

    Explorer BI4.0 SP06 does support UNV universes.

    This is an feature that is not activated by default, but you can enable it by using the instructions given here:

    1858836 – How to enable .unv support for Explorer in BI4.0 SP06

    Best regards,


    • Thanks Victor,  thought I had answered your comment.  This reminded me that the .unv support implemented in 4.1 had been backported to 4.0 SP6, but since I was working on SP6 and didn’t see it I totally forgot about it.  I added a reference to this note in the text.



  • These are issues everyone integrating BW and BO is going to face and should bear in mind, and as you’ve said not every customer is ready to jump into HANA.

    Glad you’re expanding the content with the always valuable feedback from the community.

    Really looking forward to your next blogs.

  • Eric,

    Thanks for sharing. I am looking forward to your experience with other Business Objects tools as well. I have some of my own experiences in integrating Business Objects tools on top of BW as well, and I am always wondering if the others are having similar issues. The tools are evolving rapidly and it is a great help when you specify the product versions and SP levels.