I support Web Intelligence (WebI) for a living.

I talk to WebI and it talks to me.  Mind you, it doesn't talk to just anyone - especially when it's encountering issues.  You have to know how to talk to WebI, for it to give you the good stuff - like what's really the problem.

At times, the error message WebI gives you isn't all that informative.  For example, just today the WebI document I was working on started to throw a "Failed to connect to the OLAP source. (Error: INF)" error:

WebI doesn't tell you why it failed to the OLAP source, just that it couldn't.

There's other error messages, just as unhelpful to the end user.  For example, there's a common error where the error message states "csEX".  Just that, nothing else.  Or the dreaded "WIS 30270" error.  You search the SAP Notes/KBase for "WIS 30270" and you'll get hundreds of hits, each describing a different root cause.  So which Note/KBase is the right one, the one that describes your issue?

To determine the right one, you need to find the root cause error, the actual error that was thrown that led to the more-generic error message being shown the User.

What's tricky about this is that, with BI 4.x, the root cause error likely didn't originate within the Web Intelligence Processing Server.  With the BI 4.x architecture, for improved scalability and manageability, components of the WebI document processing are done on different processes.  For example, chart generation is done by the Visualization Service housed in the Adaptive Processing Server (APS).  SAP BW BEx query data connections are done by the DSL Bridge Service also housed in the APS.  Multi-source UNX Universe data connections utilize the Data Federator Service.

So let's follow an example workflow.  You open a WebI document reporting off of SAP BW BEx query in BI launchpad.  You click the refresh button. The web browser sends this event to the web application server, the AnalyticalReporting web application handles the event, forwarding the refresh event to the Web Intelligence Processing Server.

The Web Intelligence Processing Server then determines the WebI Session associated with the open WebI document, sees that it's a refresh request for BEx query data, so sends a 'processDPCommandsEx' remote procedure method call to a DSL Bridge Service in an APS.  The DSL Bridge makes the call to the BEx query, marshals the results and sends it back to the WebI Processing Server, that composits the results in a report page, that it sends back to AnalyticalReporting that translates the XML internal format to HTML and sends it onwards to the web browser.  Phew.

But in my example WebI document above, somewhere the process broke down, and an error was thrown.  But where?

End-to-End Tracing Saved my Life

That's a pretty facetious statement.  But to be frank, the End-to-End trace tool, every day, saves me so much time and effort and makes me look pretty smart with customers. Before there was an End-to-End trace tool, trying to identify root cause within hundreds of MB's of trace log files was always a daunting task.  Much of the contents of the trace log was usually irrelevant to the issue I was investigating, and much of the time spent was manually filtering out the irrelevant.

What the End-to-End trace tool does is automatically filter out, from the trace files, all such irrelevant information.  It allows me to trace all the action taken to service a request for a WebI Session, from web browser to WebI PS to APS and back all the way to the web browser, without being bothered by any activity done on the servers not part of the workflow.

Most Web Intelligence admins are familiar with this tool - likely when you contact Support, it's the first thing we ask for.  SAP KBase 1861180 is the document that describes how to download and use the End-to-End trace when working with SAP BusinessObjects BI Platform.  I highly recommend you bookmark that KBase and keep it handy.

In fact, open up KBase 1861180.  What I'm going to describe in this blog is how I use End-to-End traces to analyze the "Failed to connect to the OLAP source. (Error: INF)" error that I got above.

To effectively use this tool, you need to know a bit of how the Web Intelligence processing workflow operates.

Preparing the System for End-to-End Tracing

The first thing you have to do is go on every machine where BI 4.x is deployed and modify the BO_Trace.ini file.  You'll note that 1861180 says to set append to false and keep_num to 20, except for WebI workflows.  For WebI, set keep_num to 50.  I asked Toby, who wrote that KBase, to put that comment in just for WebI. WebI tends to be very chatty when it comes to trace log output, and I found that for more involved issues, 20 files just weren't enough.

So make sure you edit the BO_Trace.ini appropriately:

Launching the End-to-End Tracing Tool

From 1861180, download the SAP Client Plug-In tool appropriate for the version of Internet Explorer you have installed on your client machine.  Unfortunately, currently the tool only supports Internet Explorer, and no other brand of web browser.  What the SAP Client Plug-in does is it acts as a proxy for Internet Explorer, and on all outgoing communication, injects a HTTP Header with an unique ID value.  The presence of this ID value is detected and identified by BI launchpad.  BI launchpad will then, for every action it takes to service the request associated with the ID value, increase the Trace level.  Furthermore, it sends this unique ID to any requests it makes to backend servers.  All BusinessObjects servers, including CMS, WebI Processing Servers, Adaptive Processing Servers, the File Repository Servers, etc, will also detect the presence of this ID and raise the Trace level accordingly.

So say you have the BI Platform system configured for default (Error) logging.  Any request originating from Internet Explorer launched by the SAP Client Plug-In tool will automatically raise its Trace level to HIGH for any Server servicing this request.

Unzip the SAP Client Plug-In download file and you'll see a single executable:

It's a self-extracting zip that will unzip contents and automatically launch the Plug-In:

Make sure you've closed all other instances of Internet Explorer on your client desktop, the click the "Launch" button.  This will start up Internet Explorer, and also enable to "Next Step TraceLevel" combobox and the "Start Transaction" button.  Don't click the button just yet!

The thing to know is that it'll only inject the HTTP Header with the unique ID only after you click "Start Transaction".  Before then, it won't set the trace log level.  So what you can do is, before starting transaction, do all the preliminary stuff to a point just before the action you know will trigger the error.  For my example, I log onto BI launchpad and open the problematic WebI document:

In the above screenshot, I'm at a point just before I click the WebI Refresh button, that I know will trigger the error.  I set the TraceLevel to "High" and click "Start Transaction", then I refresh the document.

Important Note:  with the WebI Web Viewer, I can wait to enable transactions till just before I take action in the viewer that triggers the error.  However, if I want to End-to-End trace workflows that involve the WebI Applet Editor/Viewer (Rich Internet Client), then I have to "Start Transaction" before the Applet is launched.  That's just the way Java applets work.  The Applet reads the web browser proxy information only at startup, and not any time later.  So you have to have the transaction started before that point.  You also need to go in Control Panel -> Java -> General -> Network Settings... and ensure the Applet is configured to use browser network proxy settings.

Once you "Start Transaction", you need to keep an eye on the SAP Client Plug-In, to make sure it's actually recording:

If the "Sent Bytes" and "Received Bytes" aren't increasing as you work with the WebI Web Viewer, then the information isn't being recorded in the traces.

After I get the "Failed to connect to the OLAP source. (Error: INF)" error, I "Stop Transaction":

You can ignore that "Settings are not valid" dialog box.

Follow instructions in KBase 1861180 and go onto all your server machines and collect the updated tracelog *.glf files.  Copy them over to your client machine for analysis.

Also, if you look in the SAPClientPlugin folder, you'll see a log folder within which is a timestamped MyBusinessTransaction folder, that contains a BusinessTransaction.xml file.  Save this file as well:

Open the BusinessTransaction.xml file using a text editor, such as Notepad++.  What you need from this file is a copy of the unique ID that the Plug-In has been injecting into the HTTP Header, identified by the "id" attribute:

Copy this id value, because you'll be using it next.

Working with the GLFViewer

KBase 1861180 has a link to the download for the GLFViewer.  The GLFViewer is that tool we use at SAP to read and analyze *.glf files.  Download the GLFViewer and start it up.

Go to File -> Open -> Add Files... and select all the *.glf files that you've collected.  Don't click "OK" just yet!  If you do, you'll get the whole contents of all the trace files, and that's not what you want.  You want to filter out any trace file entry that's not associated with the session you had with the SAP Client Plug-In.

So in the Open Files dialog box, enable the "Filter and only read matching entries:", then select Column "DSRRootContextID", then enter the unique ID value that you've copied from the BusinessTransaction.xml file:

and now click "OK".

The default configuration of GLFViewer is to not show enough columns.  Go to View -> Choose Columns... and select all columns:

Another thing I like to do is select the "ActionID" column and move it up to be second in display order.

ActionID

I'm highlighting ActionID since it's an important value.  What's an ActionID?

In the SAP BusinessObjects BI Platform 4.x system, each and every User action that leads ot a tangible change - clicking the Refresh button, opening a tree view, clicking Export, clicking the navigation button to go to the next page - generates an unique ID that all reactions resulting from the action is associated.

This unique ID is separate and different from the DSRRootContextID.  For example, when I clicked the WebI refresh button, that generated an unique ActionID.  This ActionID is then sent from the web application to the backend Web Intelligence Processing Server.  Any request that the WebI Processing Server makes to any other Service - the CMS, the FRS, the APS - also is associated with this ActionID.  All responses sent back to the web app is also associated with this ActionID.

The ActionID is the way in which we can follow the workflow associated with the User event End-to-End, from the initial request sent from the web browser, and the final response sent to the web browser.

So let's say a User request results in an error response.  To trace back the error to the root cause for that error, you would trace back, in the trace files, following the trail of the ActionID back to the first error that was thrown.

Let's do this for my example.

Web Intelligence Processing Server **Error

So let's go to the point in the workflow where we get the error message shown to the end user.

Whenever the Web Intelligence Processing Server sends an user-facing error message back to the client, it does so in XML format.  In the WebI Processing Server traces, this XML message is logged prefixed by the string "**Error".

So to find all error messages shown to the end user that passed through the Web Intelligence Processing Server, search the "Text" column for lines that contain "**Error":

It's easier if you scroll down to the very last entry in the GLFViewer, and search backwards:

Now that you have this error, read the "ActionID" associated with this log entry by going to View -> Filter:

What you've done is filtered out everything not associated with the workflow resulting in the "Failed to connect..." error.

Root Cause Error Identification

You now want to walk backwards from the **Error line thrown by the WebI Processing Server, to the actual root error.

It's easier to see this if you enable text indent, by going ot View -> Indent Text According to Scope. When you enable this, you'll see that the Text column becomes very colorful:

What the color lines define is scope.   Scope defines the depth of calls that are made internally to service a request.  For example, the WebI Processing Server invokes a request to the DSL Bridge Service.  This request is logged by a "START OUTGOING CALL" line in the WebI Processing Server trace file, that indicates that a call was made to the APS.  The APS trace file will log this request by a "START INCOMING CALL" line identifying the requester as a WebI Processing Server. The call scope has gone one deep.  When the APS returns a response to the WebI Processing Servier, it logs "END INCOMING CALL", and the WebI Processing Server logs "END OUTGOING CALL".

So from the "**Error" line identified earlier, follow the depth of call upwards until one encounters the first line in the trace with Trace column equal to Error.  You can see in the above screenshot that the root cause error actually originated in the APS, in the DSL Bridge Service.  The actual roor error is:

"Password logon no longer possible - too many failed attempts on <host>"

Someone had tried to log onto the SAP BW system using my account, but with the wrong password.

That someone was me, of course.  I had an old WebI document recurring schedule that was running, and I forgot to stop it before updating my password on my BW system.   Instant lock-up!

Stopping the schedule, logging onto the BW under another admin account, and unlocking the User resolved this issue.

Final Words

So you can see how End-to-End traces came in handy here.

You can ask yourself, "Well gee, why didn't WebI just tell the end User that the Password was invalid?"

That touches on one of the security considerations that went into the design of BI Platform.  If the system provides too much information concerning a logon failure, then a nefarious intruder can use that knowledge to try and hack into the system.  And that's not what anyone wants.

Even though this design makes troubleshooting a bit difficult, the availability of the End-to-End tool greatly mitigates the pain in getting to the root cause.

Dear All,

For all report developers and users that currently are using Desktop Intelligence and especially the well beloved Freehand SQL (FHSQL) function, there is some good news around the corner..

Today, the only option to adopt Web Intelligence, while using a Freehand SQL would be to convert your Desktop Intelligence document using a (automatically) generated Universe. This universe would then contain a derived table containing your FHSQL statement from the Desktop Intelligence document. Many users are not preferring this approach as it increases the time to enhance/maintain the SQL Statement as well as increases the complexity of the solution.

The great news I can share is that we have received approvals for the development of the FHSQL Function within Web Intelligence. With Freehand SQL in Web Intelligence it would enable you to (finally) convert your Desktop Intelligence documents with a FHSQL Data Provider to Web Intelligence. The implementation of the FHSQL function will be staged (so do not expect all features of DeskI FHSQL straight away)..

The current planned release of FHSQL for Web Intelligence will be as in the diagram below:

For the initial release of the Freehand SQL Data Provider, the following features are planned

1. Report Conversion Tool, will be enriched to convert FHSQL DeskI documents into Web Intelligence reports using the FHSQL Feature
2. Web Intelligence Documents containing FHSQL can be refreshed within the WebI Clients
3. A sample Extension Point will be provided to enable report developers to create New WebI documents containing an FHSQL data provider or remove a Freehand Data Provider (NO EDIT). This sample Extension point needs to be deployed manually after the installation of the BI4.1 Service Pack 05.
4. Support of @Variables for FHSQL Data Providers

The Freehand Extension Point sample embedded in Web Intelligence would go into the direction as shown below

After BI4.1 SP05, FHSQL will be further enhanced with additional features to close the gap between the DeskI Freehand SQL option including an integration into the UI.

Once more details become available, I will update my blog post with the latest changes.

Kind Regards

Merlijn

This document helps you to understand Dynamic Visibility of blocks on selection of Input controls with Ranking applied....

Hope it helps...

Dynamic Visibility – Input Controls with Ranking.

Imagine that, you have only one Dimension and Measure in your report but you need to display measures with different ranking options. Let us consider an example, like I require dynamic selection for Top 5, Top 10, Bottom 5 and Bottom 10 sales of products where Net Value is measure and Product is dimension object. For this to achieve, follow the below mentioned steps:

Step 1: Create a Variable called Select Rank of type Dimension with value of “ ”  and click OK

Step 2: Create Input control on SELECT RANK variable. And while adding input control pass custom values (Top 5, Top 10, Bottom 5, and Bottom 10) into it and click OK.

Step 3: Create the following Variables

Rank: = Rank ([Net Value];[Product];Top)

Top 5: = [Rank] Where ([Rank] <=5)

Top 10: = [Rank] Where ([Rank] <=10)

Bottom 5: = [Rank] Where ([Rank]>=16 And [Rank] <=20)

Bottom 10: = [Rank] Where ([Rank]>=11 And [Rank] <=20)

Step 4: Add a measure variable which will return the appropriate measure, based on user selection.

Rank Selection = If ReportFilter ([Select Rank]) ="Top 5" Then [Top 5]

Else If ReportFilter ([Select Rank]) ="Top 10" Then [Top 10]

Else If ReportFilter ([Select Rank]) ="Bottom 5" Then [Bottom 5]

Else If ReportFilter ([Select Rank]) ="Bottom 10" Then [Bottom 10]

Else [Rank]

Step 5: Consider Five Blocks. Each block holds three objects Product, Net Value and Rank. Apply filter on each block along with using Hidedimension and place them relatively in same position.

Block 1: Add filter – Rank Selection – Inlist – 1,2,3,4,5

             Hide – Hide when -- =Not([Select Rank] InList ("Top 5")).

Block 2: Add filter – Rank Selection – Inlist – 1,2,3,4,5,6,7,8,9,10

             Hide – Hide when -- =Not([Select Rank] InList ("Top 10")).

Block 3: Add filter – Rank Selection – Inlist – 1,2,3,4,5

             Hide – Hide when -- =Not([Select Rank] InList ("Bottom 5")).

Block 4: Add filter – Rank Selection – Inlist – 1,2,3,4,5,6,7,8,9,10

             Hide – Hide when -- =Not([Select Rank] InList ("Bottom 10")).

Block 5: To represent all sales, so no need of having filter

             Hide – Hide when -- =Not([Select Rank] = "")

If you want graphical presentation of data, have copy of each block and turn that copied block to chart.

Step 6:  Now your report is ready to play.

Top 5:

Bottom 5:

All Sales:

Can the collective wisdom of crowds be a successful driver for increasing the overall effectiveness of enterprise BI tools?

The answer appears to be a resounding ‘YES’ based on the successes of SAP’s Idea Place and their BI product offerings. Let’s take a look at a few facts about how SAP is doing compared to their competitors and I'll give you a few highlights about the SAP’s Idea Place as related to my favorite BI platform – SAP BusinessObjects.

Did you know…

• SAP is the worldwide leader in BI and Analytics Enterprise software
• With revenue of over $3 Billion exclusively from BI and analytics products
• Over 10,000 ideas submitted on the SAP Idea Place
• A TON of crowd-sourced ideas delivered to customers around the world and built into product offerings

Question: In an industry with an annual revenue of over $14B, how does SAP fare against the competition?

Answer: They dominate all competitors with over a 21% market share.

Question: What’s the industry?

Answer: Business Intelligence and Analytics Enterprise software

So how do they do it? Two words – SAP BusinessObjects.

Let’s drill into those numbers a little and get to the heart of the $14B number. Any idea what the largest subsegment is in the BI and Analytics segment? It’s – BI Platforms with 59% or $8.5B in annual revenue.

So, what makes SAP BusinessObjects so great?

• The BI Platform known as the BI Launch Pad– The perfect BI portal for delivering analytics to users
• The Information Design Tool (IDT) and Universe Design Tool (UDT) – The Semantic Layer rules. Need I say more?
• The Central Management Console – Security and administration at your fingertips
• Web Intelligence – The gold standard of query, analysis & reporting tools

Get Involved

• How can we as developers, analysts, consultants, users and customers be sure that our favorite (and the WORLD’S favorite) BI Platform is meeting our business needs?
• How can we become active contributors to the SAP BI ecosystem?
• How can I influence future Fix Packs, Service Packs and product releases?

Answer: That would be the SAP Idea Place– https://ideas.sap.com

Let your voice be heard and contribute to SAP’s Idea Place by voting on existing ideas, commenting on ideas to add your perspective, or submit new ideas!

So, what exactly is on the SAP Idea Place?

Products are presented in 7 different categories with the most sessions found in the Analytics grouping with 36 different sessions.

• Analytics
• Business Applications
• Cloud
• Database Technology
• Mobile
• Other

Question: What’s the most popular product session on Idea Place with nearly 20% of all Analytic ideas?

Answer: Web Intelligence

Top 5 Analytic Product Sessions

• Web Intelligence– 19.45% with 1519 ideas
• BI Platform – 11.36% with 887 ideas
• Crystal Reports – 8.49% with 663 ideas
• SAP BPC for SAP NW – 7.64%  ideas
• SAP Lumira – 4.33% with 338 ideas

Note: Disqualified from this list:

• Xcelsius - 5.11% of ideas. Since SAP recently announced that SAP Dashboards would be retired soon.
• Access Control- 5.0% of ideas. Interesting. But more governance/risk/compliance than BI.

The most popular area on SAP Idea Place is…. (drum roll)… Web Intelligence!

BI analysts, consultants, and developers alike have a ton of great ideas about Webi on the Idea Place. I also have a few submitted ideas with “a few” being... two or three… or thirty. (I hope the product developers are looking and will consider a few of them!)

What should I know about crowd-sourcing BI ideas

• It takes time.
  • You’ll have to invest some time to read through all of the ideas. Vote for the ones that interest you and that you think would deliver value to your user base.
• The ideas with the highest number of votes aren’t always the best.
  • This is simply because there are some excellent ideas that haven’t been seen by many readers yet. Comment on those ideas. Vote for them. Hopefully the product developers will give them full consideration.
• If the ideas bring value to users, the likelihood is high that they'll be delivered.
  • I have loads of confidence in the Webi (and SAP BusinessObjects) product developers and know that all ideas will receive consideration and many will be delivered.

Fact check

Below are the sites I used for financial facts used in this article:

• According to Forbes, SAP had over $3B in BI and Analytics revenue in 2013. That number is expected to increase in 2014 and the gap between SAP and it's competitors is expected to get wider.
  • http://www.forbes.com/sites/louiscolumbus/2014/04/29/2013-business-intelligence-and-analytics-market-share-update-sap-continues-market-leadership/
• Gartner agrees
  • http://www.gartner.com/newsroom/id/2723717

Thank you for reading. I’ll look for your ideas. Please vote for mine.

- Jim Brogden

http://twitter.com/jhbrogdenjr

It’s not supported, but that shouldn’t stop every BI4 Administrator from benefiting from the wonderful insight it can provide. Too often customers feel they don’t understand what the Web Intelligence server is doing; some see it as a bit of a black box. No more. The webiadmintool is a sophisticated bit of kit and worth spending a good few hours to understand how it can help you.

Here’s a summary:

1. The Profiling Dashboard will break down a single action, such as ‘opening a document’, into a large number of methods and for each its duration. Around 15 methods are shown for opening a document. This can help you identify which individual method is problematic: CMS communication, document cache, or generation of charts etc.
2. The Service LifeCycle Information shows which ‘objects’ have been left in memory and yet to be cleaned up, potentially highlighting a memory issue.
3. The Adaptive Processing Server Relationship shows, by user and document, which Web Intelligence server is using which other dependent service, like the Visualisation engine (for charts) or DSL Bridge (for the new Dimensional Semantic Layer). This means you can see inside the cluster to which user is consuming which bits of your cluster. This could be helpful to resolving a performance issue.
4. The plugins perspective allows you to “Remove a session” or “Remove all sessions using a particular document” or “Remove all session used a particular user id” should you wish to eliminate processing that is consuming too many resources.
5. The Profiling Perspective and Analysis enables you to compare two workflows with each other at a very detailed level. For example, you can compare the workflow of one user opening and interacting with a document, with the workflow for a different user. This might be helpful to isolate a performance issue, when one user ID performs slower than another.
6. The monitoring perspective enables you to visually see load (sessions, users, documents, memory usage) over time in a graph. This is ideal to validate a load test; or help you identity characteristics on the usage of your server.
7. The scheduling perspective shows you the internal tasks performed by the server, for self-maintenance and clean-up.
8. The Failover/Recovery status will help you identify the server name, user id, or document name that is frequently failing and causing a recovery to occur. Thus allowing you to take remedial steps

I’ve written a detailed wiki on the tool which I hope provides enough information for how to use the tool and the benefits for using it.

Please visit and bookmark the wiki: http://wiki.scn.sap.com/wiki/x/e5hxFw

Your feedback is, as always, most welcome. Please comment here or contact me on Twitter.

Matthew Shaw

(Twitter: @MattShaw_on_BI)

Hello everyone,

I have seen posts and gotten support cases asking about where the Island Resorts Marketing sample database has gone in BI 4.0 and 4.1 and decided to write up this quick blog that shows how to find it and publish it to BI.

On the server go to <Install Dir>\Program Files (x86)\SAP BusinessObjects\SAP BusinessObjects Enterprise XI 4.0\Samples\webi

You will see the club.mdb and Island Resorts Marketing.unv.

Open the Universe Design Tool and Open the Island Resorts Marketing.unv

Once it is open, click on File|Parameters

Under the Connection drop down, select New

Create a new Secured Connection, naming it whatever you want and click Next

Select Microsoft|MS Access 2007|ODBC Drivers then Next

On the Data Source Name drop down, select club-webi, which is a DSN created when you install the product. If it is not created you will need to create a DSN yourself to the Island Resorts mdb (for example on your desktop machine)

Click on Test Connection

Click OK then Next

Click Next again

Click Finish.

Verify your new connection is being used by the universe

Click Ok

Select View|Refresh Structure

When prompted to refresh out of date tables, click OK

Then click Ok again when it states they are updated.

Click Save

Click File|Export

Browse to the location in BI that you want to store the universe

Click Ok then Ok again to export it

You can now use it to create sample reports!

That's it!

You can now use the Island Resorts Marketing universe for your testing needs!

Hope this helps!

Jason

Hi,

You might be interested to know that several SAP BusinessObjects development guys (Product owners, Domain Architects, Product Experts) are taking part in SAP TechEd && d-code in Berlin next week. This will be a great opportunity to exchange ideas, best practices and roadmaps with our customer, partners and internal colleagues.

My colleague Bruno Louifi, Christian Ah-Soon and myself will be presenting a hands-on session on a fresh Web Intelligence feature called "Extension Points" (session ID: EA267 - Tuesday 11th Nov at 16.45 H3 Hall 6.2).

The idea behind Extensions Points is to "programmatically" extend Web Intelligence 4.1 capabilities in order to enable new business scenarios that are not available out of the box. Examples of extensions include connecting a WebI report to Google Maps, Predictive Analytics, creating annotations, writing back to a database...All using Javascript goodness and the fresh Web Intelligence RESTFul SDK.

Bruno, Christian and I will be there to guide you through the exercises and to assess the feasibility of your project ideas.

In summary, during this hands-on you will:

- Configure the Eclipse environment and create your WebI Extension OSGI bundle (e.g., creating hyperlinks on the left panel to interact with the report)

- Configure Deploy Test and debug your extension in order to make it available in your Webi Report viewer.

- Deploy some samples for you to get a taste of this powerful feature.

The target audience would be customer/partners/consultants with some dev skills (Java, JS) and with some basic knowledge on Web Intelligence workflows.

We had some positive feedback on this session which was first ran in Las Vegas, hope you'll agree and share the excitement, which proves that we are still continuously innovating in our good old Webi !

See you in Berlin!

Even though the title tells about merging the two charts, in reality this is not possible in Webi . The idea is to trick WebI by putting the chart on top of each other in such a way that they appear to be merged . The idea for this blog came from the following query posted in the forum.

http://scn.sap.com/thread/3652142

Requirement

To plot a Cross tab into a Bar/Line chart in such a way that Measure 2 forms a bar chart for all Zones for a particular Date and Measure 1 forms a Line chart for each Zones for a particular Date. This is the tricky part because we need 2 measures to behave in 2 different way in the same chart. To my knowledge this cannot be achieved in a single chart and hence we will plot them separately and then merge them. 

Preparing the Dataset - First we will create a data set for plotting our graph

Below is a Crosstab table which has Date along the Vertical Axis and Zone along the Horizontal Axis and a Measure Amount on the Body Axis

Also will create a variable Average which will be Measure 2 and add a column at the end of the table. So, that’s how the table and our data objects looks like

Plotting the Chart

Now we will try to plot this data into a Bar/Line. We will select the Combined Column Line chart. X Axis will be the Date. Value Axis will be the Average(Bar) and Amount(Line). And since we need the Line chart for each Zone, we will select the Region Color as Zone.

Right Click Chart -> Format Chart and make sure you select the below option from the Measure properties. So that Average is Bar and Amount is Line.

And we get the below Chart. The issue with this is; the Average also get separated for each Zone which is not something that we want. We want the Region Color to be applied only on Amount and not on Average.

Workaround - Following are the steps taken for the workaround.

       1. Create 2 separate charts

       2. Format the charts

       3. Place the chart on top of each other

Creating a Bar Chart

Create a bar chart for the Average as shown below. We will not select Region Color since we don’t need this to separate into the Zones.

This is how the chart looks like and exactly what we want.

Creating a Line Chart

Create the Line Chart. Here we will select the Region Color since we want to separate the Amount into zones

This is how the chart looks like and exactly what we want

Formatting the Bar Chart

First we need to format the base chart (BAR) as shown below.Right Click on the Bar Chart - > Select Chart Properties

Goto the Value Axis tab and select the Design option. This is to fix the Y-Axis. Make sure to select the Maximum value in such a way that it is higher than all measure values in the table

Scroll down below and hide the text of the Value Axis by selecting the color as white. Don’t uncheck the Visible checkbox else the chart will get dislocated.

Similarly goto the Title option and hide the text as well

Repeat the same for Category Axis and hide the text of the Title and Design Option as shown below

Hide the legend Title

Change the position of the Legend value

And the Bar Chart will look like as below

Formatting the Line Chart

The first and most important thing is the make the Y Axis of line chart same as Bar Chart. This is because when we put the chart on top of each other, we want the axis to be same. For formatting right click on the Line chart -> Select Chart Formatting

Next we remove the legend title from the Legend -> Title

Goto the Value Axis -> Title and give a custom title name

The next important part, is to make Line Chart transparent so that when it is placed on top of the Bar Chart, we can see both the chart. For this reduce the RGBA Color to 46%.

We won’t touch the Category Axis because we want to see the Axis details when we plot the charts. One more setting that we would do is make the background white for the line chart.

Now our Line chart will look like as below

Merging the Charts

Now the last part of this trick, make the relative position of the Line Chart same as that of the Bar chart

And we get the Chart as shown below which matches with our requirement

In support, we get a lot of questions and feature requests.

One request that has passed my desk a few times is the ability for a webi document to use an unmanaged Excel spreadsheet as a data source.

You might not be familiar with the term "Unmanaged" so I'll explain a little.

BI 4.1 SP02 introduced a new feature that allowed you to base a Webi Document off of an Excel spreadsheet that was published/saved to the BI Launchpad environment. This is referred to as a "Managed" object within the BI Platform.  A Managed object is something that is imported into the BI repository and that is managed by the CMS.   This is great if your data is static in the Excel sheet but lacks functionality if the data is being updated frequently.

So, the question has come to me a few times now:  Can we have a Webi document use an Excel data source that is "Unmanaged"?  Meaning it is not published within the BI Platform and exists on a network share.

I didn't know if this was possible at first but after about 10 minutes of playing around, I was easily able to accomplish this using the below steps.

Keep in mind this is just a simple walk-through and doesn't cover all scenarios, but this should give you the basics on creating a Webi document that will refresh against an Excel data source that exists outside the BI Platform environment.

So, here we go:

1. First, place your Excel file on a network share.  You can see below, mine is on a share called:  \\bipw08r2\dataSource\data.xlsx
   
2. The data is simple in this example.  3 columns and 4 rows of data to start. 
   
3. OpenWebi Rich Client (WRC) and create a new doc.  Choose Excel as the data source
   
4. Type in your network share name in the File Name box.  Select your xlsx sheet as I did below
   
5. The Custom Data Provider - Excel dialog will allow you to choose a source path, sheet, range, etc... 
   
6. After you select your options, you will see the Query Panel.  Note the Source Path shows the network share and the Result Objects/Data Sample reflect what is in my sample xlsx.
   
7. I click Run Query and see my Webi document with the data from my spreadsheet.
8. To ensure my Webi Doc is refreshing, I am add a new row to my spreadsheet and save it to my network share location
   
9. I then refresh my Webi Document and see the new row appear.
   
10. I save my Document out to my BI Platform using the "Save to Enterprise" option under the Save icon in WRC.
11. Then I open up a browser and login to BI Launchpad.  I navigate to the folder where I saved my document.
12. To test that my BI Launchpad report is refreshing against my network data source, I update the Excel spreadsheet and add a row. 
    
13. Then I refresh my document in BI Launchpad to ensure the new row is there
    
14. I now have a report that will use an Excel sheet off of a network share to get data.   This gives me some more flexibility around my data for my report.

There are a few things that you will want to watch out for with this option.  I've documented the ones that come to mind below:

• In most cases, the SIA (Server Intelligence Agent) will need to run as a network account that has rights to this network location.
• You will need to ensure you have an Adaptive Processing Server that has the Excel Data Access Service running on it (and enabled)
• Only the Applet Interface and WRC interface support modifying/creating Excel data sources for Webi Docs
• The "refreshable" option will need to be selected for the Data Provider in order to refresh a document
• The "editable" option has been removed in later versions of BI 4.1 because it was not needed for Excel data sources and did not do anything except cause confusion to end users.

That's it!  I'd love to hear from anyone using this in the real world.  Are there any gotchas?

Thanks!

Jb

As many have heard, free-hand SQL is making its long awaited debut in BI4.1, in the form of SP05. As an engineer in the Webi Product Support team, I was pretty excited to have access to an early release of the product, so I decided to test it out.

I've been with the company since the BO6.5 days and like many of our loyal customers, I liked Deski. It was feature rich and many times, for whatever reason, worked faster than Webi. One of the features it had was the ability to create reports using free-hand SQL.

One of the many benefits of using free-hand SQL was the fact that a Universe wasn't needed. All a report designer needed was a connection to the reporting database and their SQL skills. Custom SQL was directly entered into a dialog box in Deski and data was returned. No muss, no fuss.

Because of the deprecation of Deski in BI4.0, and free-hand SQL not yet a part of the features in Webi, many of our customers refrained from upgrading. Now I know upgrading isn't fun, especially a robust product like SAP BusinessObjects Enterprise, but at least we can cross off the inability to use reports created using free-hand SQL as a reason. I had limited time with the pre-release of Support Package 05, so I was able to do a simple test. I've documented my experience and recommend extensive testing for anyone interested.

*Note* You can't actually refresh a migrated Deski report in BI4.1. I just want to make it clear that Deski isn't back. However, with Support Package 05, we now have the ability to convert FHSQL Deski reports into Webi reports.

Step 1: Create a Deski report in XI3.1 with FHSQL. Ahh the good 'ol days. Unfortunately, I realized my SQL skills are a little rusty. So I created a very simple one. I chose Free-hand SQL as my datasource.

Step 2: Create a connection to the reporting DB.

Step 3: Enter the custom SQL in the dialog box and click Run. Voila! Save and export the report to the repository. I created a second report with an @variable defined. In my case a prompt for Country.

Step 4: Create a .biar file with my reports. The Source system is XI3.1 and the destination is a .biar file. Because this was a simple test, the things I brought over are just the 2 reports.

Step 5: Bring the *.biar file to the BI4.1.5 environment and use Upgrade Management Tool to import the FHSQL reports. I'm sure Promotion Management Tool would work too, but because this was a simple test with just 2 reports, UMT was used.

Step 6: Once the Deski FHSQL reports are imported, use the Report Conversion Tool to convert them to Webi

Step 7: Confirm that the Webi reports have been created and published.

Step 8: Moment of truth! I refreshed the Webi report with the prompt to make sure it worked. I initially created the report with "US" as a prompt value (shown). I changed it to France to make sure that it without a doubt refreshed!

I was quite happy and impressed with how well it worked. Don't get me wrong, I ran into a few issues which I will talk about in a separate blog, but once I overcame those challenges, things worked beautifully. Back in the Deski days, Report Conversion wasn't a trouble-free task. Regular reports didn't convert for a number or reasons let alone FHSQL reports. SAP has brilliant customers that do amazing things with our product. Things we maybe didn't think of doing, or testing. I remind you again that even though the above test worked awesomely, they really were simple reports. So if you're considering upgrading now that Support Package 05 is available, TEST, TEST, and TEST some more. Good luck!

Hi everyone!

Some of you may have seen my wiki on Web Intelligence GUI Customization that came out with BI 4.1.

In BI 4.1 SP05 we have added onto this functionality by allowing you to do this customization on a per folder basis instead of just globally.

Customizing the interface is still done at the group level in the CMC by just right clicking on the group name and select Customization.

You will now see the following screen:

The basics for setting up the User Interface Elements, Features, and Extension are still the same as they have been.  You can click the link above to view the wiki that explains the features.

The new functionality is being able to set this up on a per folder basis.

The Customization folders section contains a folder named Default Folders, which is used to define default customization.

You can choose folders for which you want to apply specific customization by clicking the Add Folder button.

You can avoid redefining the same customization for other folders by copying customization from one folder to another by using Duplicate Customization and Paste Customization options from the drop-down list. If you want to remove customization for a specific folder you can click on the Remove folder option from the drop-down list.

The following rules are used to define customizations to apply to a user:

• If the user belongs to different groups, only the customization defined to the group whose ID is lower applies. The customization defined for the other groups containing the user does not apply.
• For nested folder structure, the immediate parent folder of the document that has been added in the list of customized folders defines customizations for the document for user interface elements, features, and extensions.
• The customization defined for Default Folders apply for the documents stored in Personal Documents and Inboxes, and for documents for which the parent folder is not customized.
• The customization defined for user interface elements have priority over customization defined for features as features is only a shortcut to enable all user interface elements.

I hope this helps!

Please leave any comments or suggestions below.

Best Regards,

Jason

Have you ever used multiple selection type Input Controls in your Web Intelligence report?

Did you ever find it frustrating that WebI refreshes the data view before you were done selecting all of the values you wanted in that input control?

Starting in SP05 for BI 4.1, WebI has a new feature to add an "OK" button for "Multiple Selections" type Input Controls for Check Box and List Box.

It can be enabled or disabled in existing Check Box Input Controls by using the edit option in Design Mode or when first adding the Input Control.

The OK button is enabled automatically for the List Box Input Control and cannot be disabled.

How does it work?

Let's take a look at an example using the eFashion sample data source.

We've added an Input Control of Check Box type for Stores to allow users to narrow down the Sales Revenue they see for the selected stores only.

The existing behavior is that WebI will refresh the view of the report to only show the data of the selected check box AS IT IS SELECTED.  If you want to select 4 stores, the view will update 4 times after each check is made.

With the new addition of the OK button, the view will not update until the user clicks the OK button.  This way, your user can select all 4 stores, then click "OK" to have the view refreshed.

This can be a huge time saver when your report contains a lot of data.

What does it look like?

List Box Control Example:

Check Box Control Example:

How do you enable the OK button for the Check Box Input Control?

Existing Input Control:

1. Open the report in Edit/Design Mode
2. Select the Input Control menu on the left hand side
3. Hover over the input control to display the interactive icons
4. Click the wrench icon to edit the Input Control
5. Click "Yes" next to "Display selection OK button"

New Input Control

1. When adding a New Input Control, only the "Multiple Selections" types will have the OK button and only the Check box type will allow you to enable or disable the display of the OK button
2. The option to enable it is the same as above.  Under the Input Control Properties, you will check the box next to "Yes" for "Display selection OK button":

Let us know what you think!

REFERENCES:

For more information on what's new in BI 4.1 SP05, check out the SP05 What's New Guide found on the SAP Help Portal.

In my previous blog post, I talked about my experience with converting free-hand SQL Deski reports in BI4.1 SP05. The resulting Webi reports worked flawlessly. That’s not to say I didn’t run into any issues.

The fact that this feature is available is great. There are, however, a few limitations.

We’ll start with what the feature allows you to do:

1. Using Report Conversion Tool, you can now convert a Deski report created using FHSQL into a Webi report.
2. The FHSQL data provider can contain items like prompts, variables, etc… There aren’t any restrictions.
3. The resulting Webi report can be refreshed successfully.

Here are a few limitations and gotchas:

1. The FHSQL feature allows you to convert a FHSQL Deski report to Webi, but you CANNOT edit the SQL in Webi.
2. At the moment, without the extension point, you cannot create new FHSQL Webi reports.
3. Please make sure that the database drivers used to create the FHSQL connection in XI3.x are installed (and supported) in the BI4.1.5 environment.

Troubleshooting Tips:

During my initial tests, I ran into a major snag. My Deski reports just would not convert. They would fail with an error: “Failed: Error on Migration Engine” & “Not Converted : FHSQL1 - .wid document could not be generated due to unknown reason”. Here are things I tried to resolve my issue.

1. Client side tracing. I followed KB: 1586166 - How to enable tracing for BI4.x client applications to turn on client side tracing while trying to convert the reports.
2. I tried using Conversion Audit option in RCT
3. I thought I needed other things in the biar file, so imported things like Repository Objects, Universes, etc…
4. I played around with the settings in RCT like Read Cell as Text
5. I retraced my steps from creating the report in XI3.1 to converting it in BI4.1.5

Client Side Tracing:

This KB 1586166- How to enable tracing for BI4.x client applications is very useful. I copied the BO_Trace.ini file from the dir: <\SAP BusinessObjects\SAP BusinessObjects Enterprise XI 4.0\conf>

I followed Option 2: Tracing via environment variable. I initially just enabled regular tracing:

//==========================================================

//Trace Syntax and Setting

//

//uncomment the lines below to enable tracing via ini file.

//This will override the trace setting via command line

//

//==========================================================

sap_trace_level = trace_error;

sap_log_level = log_error;

//keep = false;

//size = 100 * 1000;

But I found that this didn’t get me the information I wanted. So I changed the following lines:

sap_trace_level = trace_error;

sap_log_level = log_error;

to

sap_trace_level = trace_debug;

sap_log_level = log_info;

This gave me verbose trace logs. Now don’t be surprised if there are multiple files generated. I saw the following:

All these files are important. The Report Conversion Tool calls are in the TraceLog file, but it’s a java process hence the java logs, and it also invokes the Universe Designer Tool, we believe, for connectivity information. If you run into any problems with RCT, and the issue isn’t obvious, then I highly recommend verbose tracing and collecting all these logs.

Now at the time of writing this blog, the reason for my issue was not apparent in the logs. We saw error messages, but it didn’t scream “I did it”. We have an open dialog with our development team about possibly improving our error messages.

Just recently we also discovered that the RCT might also be requesting information from the APS. I figured out the reason why my reports were not converting, which I will talk about, but I will update this blog after I run a new trace and this time include the APS in the tracing process. Stay tuned!

Conversion Tool Audit

Since I was not having any luck with trace logs, and my patience was wearing thin, I decided to try a different avenue of troubleshooting. My server has Auditing enabled so I thought I’d audit the RCT activity to see if it clearly states why my reports were not converting.

This KB is what I used as a guide: 1846268 - How to enable Report Conversion Tool Audit in BI 4.0

Things to keep in mind:

1. You need to enable auditing and have an auditing db.
2. The KB talks about using Universe Design Tool, so make sure you have 32bit drivers for your DB installed.
3. If you update the universe before you run the RCT, you will not see the appropriate table. Run RCT once with Auditing enabled(image below) and then you will see the table in UDT.
4. Don’t forget to enable the Audit settings in RCT as shown below every time you run RCT. I wish it stayed on all the time, but I can see how that’s not always appropriate or necessary.

Here is what the Report Conversion Tool Report looks like after converting 2 bad reports and 2 good reports.

Let's take a look at the "Not Converted" report tab.

While the report gives good information, it didn't tell me exactly why my reports were failing to convert.

Migrating other objects

I tried migrating other objects like Repository Objects, Universes, Rights, etc... Because my test was really simple, I really didn't need anything but the reports. I can understand that as a customer, more likely than not, you would do a full system migration. Please note that while that's perfectly fine, if you're testing reports, using the method I mentioned in my previous blog is just fine.

Changing the settings in RCT

At this point I was grasping at straws, so I thought I'd change the settings in RCT. This was pointless as all I really needed was to check the check box that read: Convert documents containing free-hand SQL / stored procedures.

Retraced my steps

Finally I decided to retrace my steps. In doing so, I realized my mistake. I had created the FHSQL Deski report using a MS Access 2003 ODBC driver. In my BI4.1.5 environment, I didn't have that driver installed. So I created a new FHSQL Deski report (although updating the connection would have worked too) using MS Access 2007 ODBC driver, migrated the reports and converted them SUCCESSFULLY!

It took some trial and error, but we figured it out. In the process we were able to identify areas that need improvement, we defined a few possible troubleshooting techniques, and actually celebrated the fact that once it worked, it was awesome.

First look at FHSQL reports in BI4.1

For the latest information about Web Intelligence, check out the WebI Bulletin

Charlotte from the Documentation team sent word a few weeks ago she's been working on a new doc piece for Web Intelligence - a Getting Started Guide.  I think it's pretty nifty and extremely useful.

If you belong to an organization that uses Web Intelligence, keep a reference to the new doc handy, cause you'll use it.

One of the compelling reasons that keeps Web Intelligence an extremely popular reporting tool is how quickly users can create compelling reports.  Way faster than other tools (a quote heard at last year's ASUG BI User Conference: "Hey, I can use WebI and go for lunch, or use another tool and try and eat at my desk").

How to get the users started?

SAP provides training and courses (https://training.sap.com) for the beginner to advanced.  But at times what's needed is a quick intro that covers the major features - creating a query, table and charts - is what fits the bill.

The Getting Started Guide is just that:

To find it, go to SAP BusinessObjects Web Intelligence 4.1 – SAP Help Portal Page and look for the Getting Started guide in the End Users Guide section.

It walks the user through the creation of a "real" report using the eFashion sample Universe we provide with the product.  It's been targeted for BI 4.1 SP05 Product release, so it's now public, but I'll comment that the walkthrough applies even for older versions of BI 4.x.

What makes the new documentation that much more compelling is that Leah created a companion video for it:

(click above image or go directly to Scenario - Creating Reports with Tables, Charts and Images: Web Intelligence 4.1 SP5 - YouTube)

If you have any users who want to get started on Web Intelligence, do foward them the link to the Getting Started guide - it will walk the user through the creation of their very first WebI doc.

While you're at SAP BusinessObjects Web Intelligence 4.1 – SAP Help Portal Page, check out the new docs for BI 4.1 SP05!  You'll notice that our Documentation team has made changes with newer SP's to make it easier for users and administrators to find the right info.

Previously, the docs were fragmented.  We had one Web Intelligence Rich Client User Guide, and another for using WebI with BW/BEx queries.  There were considerable overlap in content, so a decision was made to unify the documentation in one Web Intelligence User Guide (the admin information remains in the BI Platform Admin Guide).

We also have a new initiative going where we can foward any feedback you have concerning the BI documentation to our documentation authors.  So if you find anything amiss with our BI docs, please do forward the comment if you happen to have a Support ticket open with SAP.

Hi Everyone,

In my first blog post I´ll like to share a trick to set a page break into a report.

Sometimes, in order to better format a Report, one needs to add a page break on it.

Consider a requirement that is to render a report with charts and tables but the user wants the tables and charts to be shown in different pages. Firstly the tables should be shown and in a following page, the charts.

Create your table and give it a name, lets say MainTable

Create a blank cell, named PageBreak. After that, on Cell Properties (Format Cell->Layout), check Vertical -> Start on a new page) like the picture below.

Align PageBreakl by the bottom of MainTable

After that, align the charts by the bottom of BlankCell

Align all the charts based on this cell, Page Break.

Hope this is clear, otherwise, feel free to contact me throug rogerplank1234@yahoo.com    

Cheers,

Rogerio

Hello Everyone,

Today, when a webi report based off BEX Query is refreshed, the only way to enter the prompt is to select it from the List of values.  This has changed as of last week with the release of BI 4.1 SP 05.

I am delighted to share with you all that a new manual entry field has been added in the WebI prompt UI. This new functionality will allow users to enter the prompt value manually instead of having to select it from the LOV.  This functionality is in addition to the already existing option where users can select the value from LOV.

The below screen shows the multiple values selected for the prompt (key values entered manually):

The keys can be entered in different ways. The list of supported input formats are as follow:

1. Single Member: 1

When using Selection Option Variable based prompts, the prompt can be shown as multi-value prompt instead of an interval prompt by making the following entry in the configuration.  The input format would be as follow:

1. Interval: 1-5 (do include the spaces)
2. Expression with operators: >4, >=4, <4, <=4
3. Exclusions: !5 (any key but 5)
4. Combinations: 1-10;!5 (all the keys from 1 to 10 except 5).

You can also use pattern matching feature to speed up the selection of a set of keys. Supported pattern matching could be something similar to:

• *1 : Retrieve all keys that end in 1 ( for example:  01, 11, 21)
• *1* : Retrieve all keys that include 1 (for example: 01, 10, 11, 12, 13, 21, 100)

In the next blog, I will update you with the steps on how to configure settings in order to enable multi value for Selection Option Variable based prompts. 

Hope this helps.

Regards,

Ejaz

Ever try to design a report using a large dataset?  Doing this can be a performance nightmare as you may not have applied the filters you wanted or the query to the database may take a long time to complete.  So how do we improve this?  One feature that exists in Web Intelligence to help with this is called Data Sampling.  This feature already existed in the product for many relational databases through the unx format and in SP05 support for this feature was added for relational HANA connectivity.

I had never used this functionality, so the first question for me was, where do I go to turn it on.  The answer is, within the query panel by pressing on the Query Properties button as seen below:

Pressing this button brings up the Query Properties screen where there is a section called "Sample" which has two properties, "Sample Result Set" and "Fixed" as seen here:

You may notice that the "Fixed" property is greyed out, this is because currently this functionality is not available for HANA.  What the fixed functionality provides is so that each refresh will return the same data each time.

My next question was, what does this sample result set actually do and how does it work.  I referred to the Web Intelligence User Guide which in section 2.1.3.7.3 tells us that it applies a database level restriction on the query.  So what does this really mean.  The first thought was, this must mean it is part of the SQL that is being generated, so I clicked on the view SQL button within the Query Panel, as seen here:

This brings up the SQL Statement that will be run.  In my case, this is what I saw:

What we see is that part of the SELECT statement, we have added TOP 500, as that was my Sample Result Size I had set earlier.  Additionally, at the end of the SQL statement, we see ORDER BY RAND().  If you set this property when you first see the query panel when creating a new document, you never need to wait for a large query to return, and you can come back anytime to uncheck or recheck this property as needed.

How does this differ from the "Max rows retrieved" setting?  This setting queries the database for all values, but then truncates the result that is actually displayed and is when you will see the partial results warning next to both the refresh status at the bottom of the report, as well as in the data tab.  To access the data tab, you click the button at the top right of the viewer as seen here:

On this screen we can see the following warnings related to these two settings which let you know if there are partial results, sampled results, or both.  The both scenario is when the sampled result size is greater than the max rows size.  Below are images for what you see for each of these scenarios.

Partial Results:

Sample Results:

Both:

I personally see this as a simple and great way to speed up the initial designing of reports.  Of course we would want to remove this setting before doing any final testing to make sure our system would be sized to handle the actual data being returned as well as verifying the calculations with the larger data are what you expect.

I hope that this blog was helpful and that you will try this functionality out when next designing a document.

Also, please bookmark our Webi Bulletin as it is a great page that is updated regularly with information on the Web Intelligence product.

Hi Team ,

It's simple trick to display totals of stacked column like below

On top

at bottom (along with category axis)

How to do !

On Top

I have Business area and fiscal period wise Revenue stocked column chart.

1.Create a stacked column chart with Business Area in category axis ,Revenue in Values axis  and Region Type : Fiscal Period

2.Now we need totals to display on top .Create a detail variable to get Business are wise totals .

Total =FormatNumber([Capital  Cost FC] In ([Business area]);"#")

Note : If you have big numbers ,then it is better to show them in Millions (if Value)/MT (if Quantity) .So that we can adjust totals to look like for each column.

3.Create a variable with dummy measure .This is to have measure against Detail total variable

Dummy =100

4.Create a column chart (any chart ) with Total variable in X- axis , Dummy measure variable in Y-axis.Fiscal period in region type to have context (if any to measure)

5.Show only category axis only .Unchecked value axis, title,legend ,..

6.select palette for chart : create a custom palette with all white colors

This is to don't show columns in chart.Format column chart category axis to show numbers in bold color .Decrease height of column chart to get the category axis and some space for columns .By doing this we should get values like below.

7.This is important part .Align this column chart on top of stacked chart so that it will appear as total of the column.

Cons :

* we need to have another chart , it will heavy our report .

* we should be very concise on formatting and aligning.

at bottom

Assign Total -detailed variable in category axis .Insert a blank cell , name it Totals and align it as to represent category axis total .

Cons:

Totals are shown at category axis itself .We should inform to user .

Hope somebody will find it helpful.

We often get questions regarding strange calculation results when using decimal numbers in Web Intelligence.

Typically:

• Why don't I get 0 when I do A + B - C, although C is supposed to be equal to A + B?
• Why does the result of this running sum is different when I change the sorting of my table?
• Etc.

These strange calculation results are due to the fact that, internally, Web Intelligence represents decimal numbers using the 64-bit precision version of the Floating Point data type defined by the IEEE 754 standard.

A detailed explanation of that standard can be found in the following Wikipedia page: http://en.wikipedia.org/wiki/IEEE_floating_point.

With the IEEE 754 standard on floating point arithmetic, decimal numbers cannot be perfectly represented in binary form, since not all their digits can be trusted. In fact, the number of digits which can be trusted actually depends on the size of the representation. With the 64-bit representation used in Web Intelligence, the number of bits used for stocking the data is 53 and, as explained in the above Wikipedia page, the number of digits which can be trusted is: log(2^53), rounded to 15.

IMPORTANT TO REMEMBER: In Web Intelligence, the maximum number of digits which can be trusted in a decimal number is 15 and that takes into account digits both before and after the decimal mark.

Example:

• 100,000,000,000 + 0.001 will correctly result in 100,000,000,000.0010 because this decimal number only requires 15 trusted digits to be represented (12 digits before the decimal mark and 3 after).
• 100,000,000,000 + 0.0001 will wrongly result in 100,000,000,000.0000 because the correct result would require 16 trusted digits to be represented (12 before the decimal mark and 4 after)
  

With this limitation in mind, summing decimal numbers might not always give the expected result, especially if the expected result is 0...!

For example, summing 19000.2, 0.123, -9100.3, -0.000000000002543, etc. will allow for only 10 digits after the decimal mark to be trusted. Indeed, 5 digits before the decimal mark are already used by the largest number 19000.2. Consequently, if the result of that sum is something like: -0.000000000013327, it will be indistinguishable from 0, by the IEEE 754 standard.

Therefore, if a 0 result from a decimal numbers calculation is used as a condition for further processing in a Web Intelligence document, it is highly recommended to convert the decimal numbers into integers before that condition is evaluated. This rounding operation can be done with the Round(number; round_level) formula, using 0 for the round_level parameter.

Now, why is changing the sorting of a table might also change a decimal calculation result?

Again, this is due to the IEEE 754 floating point representation. Because of the limited number of trusted digits in this representation, there will be a truncation error when, at some point during the calculation, the precision is exceeded. This error will then be propagated along the remaining of the calculation, leading to different results according to the way the values are sorted.

I hope these "strange calculation results" will be clearer now!

If not, please do not hesitate to ask questions below.

Best regards,

Pascal.

Hello Everyone,

In my last blog, I mentioned about the supported input formats for selection option BW variable for the newly introduced feature of manually entering BW prompts.

I tested the new feature further in webi rich client and after some effort was able to make it work.

Here is the step by step instruction on how to make this work.

1. I started testing this feature by first creating a BW Query with a Selection option variable prompt as shown in the screenshot below:

2. This is the most important step. Add a registry key for a client machine running webi rich client at the location below:

The Value Name is the next number in sequence. Depending on your configuration, it may be different then what you see below. The Value Data for the key is “–Dsap.sl.bics.variableComplexSelectionMapping=multivalue”.

3. I then created a webi report based off the BW query using Webi rich client. I tried using the input formats as mentioned in the last blog (intervals, Expression with operator and so on).

4.  I added multiple prompt values for Country Key  as shown below:

5. Ran the report and checked the results:

Above dataset in the report shows that data brought back consist of countries with key value 1,2,3,4,6,25 and 33 excluding 5 (as NOT INCLUDING 5 was one of the conditions.)

I hope this is helpful.

Regards,

Ejaz