TekLink

All reports in SAP business warehouse get their data from Infoproviders. InfoProviders are objects that store data in SAP business warehouse. These Info providers can be objects that physically contain data or logical views, whose data is stored in another system or other info provider.

There are many types of InfoProvider objects in SAP BW. These include InfoCubes, operational data stores (ODS) objects, InfoSets, remote cubes, SAP remote cubes, Virtual InfoCubes with services and multi-providers.

InfoCube

An InfoCube (basic cube) is made up of several related tables that organise the info objects contained within according to the InfoCubes Star schema. The central table in the InfoCube is called the fact table. The fact table contains the key figures and dimension ID’s that link to the dimension tables that surround the fact table, making up the Star schema. The dimension tables themselves don’t actually store the characteristic data or data keys, but store another numerical ID, known as a master data ID which is automatically generated by SAP BW.

The master data table (SID table) is used to create the link between the master data ID and the value of the characteristic.

Operational data store

Operational data store (ODS) objects are basically transparent tables do contain the key fields and data fields that make up this type of info provider. Even though, in BW they go by the name ODS objects they are basically the same standard tables that hold data in SAP R/3.

There are a few limitations with this type of info provider. The main one of which is that they are limited to a maximum of 16 key fields. As ODS objects, are just basic flat tables, all the fields they contain (characteristics and key figures) are held in each data record, which means your records can become a considerable length.

Although ODS objects have limitations, they are ideal objects to store data at a document level.

An operational data store object stores consolidated and cleansed data (transaction or master data) and data fields that can also be key figures and characteristics, which differ from a basic cube.

And ODS object describes a consolidated data set from one or more Info sources. You can evaluate the dataset with a BEx query. And ODS object contains a key such as a document number or item and data fields that can also contain character fields such as customer as a key figure. The data of an ODS object can be updated with the Delta update into InfoCube or additional ODS objects in the same system or across systems. Contrary to multi-dimensional data storage with InfoCubes. The data of ODS objects is stored in a flat database tables, unlike basic cubes. ODS objects consists of three flat tables.

• Activation queue (initial table of ODS objects).

• New data is stored in this table before it is activated. Its structure is similar to that of a PSA table that keys built from the request, data package and data record number. After all the requests in the activation queue have been successfully activated. They are deleted from the activation queue.

• This table stores, the current state of the data. This table has a semantic key such as an order number or item that the model can define. Reporting draws upon this table. If the connected data targets are supplied in the full update method of updating, the data targets are updated with the active data from this table.

• During an activation room, the modifications are stored in the change log. The change log therefore contains all the complete (activation) history of the modifications across the contents of the change log and not automatically deleted. If the connected data targets are supplied from the ODS object in a Delta process, the data targets are updated from the change log. The change log is a PSA table and can be maintained in the PSA tree of the administration workbench. Accordingly, the change log has a technical key derived from the request, data package, and data record number.

The new state of the data is written in parallel into the change log and into the table with the active data. Note the following types of ODS objects:

• Standard ODS object. These are created in the administration workbench modelling SAP, under info provider. Create them under the required InfoProvider. You can also edit ODS objects using transaction RSDODS.

To Be Continued …

Technorati Tags: Infoproviders, SAP, BW, InfoCubes, ODS, InfoSets, remote cubes, SAP remote cubes, Virtual InfoCubes, multi-providers, Star schema, fact table, dimension tables, SID, characteristics, key figures, PSA

Whilst working with BW BEx users, I often get asked to explain what a key date is and how it affects query results. It’s understandable that this can sometimes be a little confusing to users as you really need to know a little bit about how your BW handles Master Data within your data model star schema.

Characteristics can contain attributes, which help describe in detail information about the characteristic itself.

• For example The 0EMPLOYEE characteristic holds the value of an employee’s personnel number. This is great but it would be better if we new more information about this employee. This is where attributes come in. We can assign attributes to the 0EMPLOYEE characteristic that describe more information about the employee such as Job, Position, Gender, Personnel Area etc…

When attributes are assigned to a characteristic, we can add these attributes to our reports in the BEx Analyzer. When we run the report, the BW system reads the attribute values that correspond to the characteristic and displays them in our query results.

Woohoo, I hear you cry, but what about the Key Date?

When attributes are assigned to characteristics, there is an option to specify that the attributes are Time Dependant. This means that instead of the BW system reading a single attribute value that corresponds to the characteristic, the system has to use dates to workout if the attribute value is valid for the reporting time period the user has chosen when running the report. This is made possible because when Time Dependant data is loaded into a BW system for a characteristic, a start-date and end-date is loaded with it identifying when an attribute is valid.

Now when a query is executed that displays attribute information; the BW system needs to know what date to use when searching for the relevant attribute information. This is where the Key Date comes in. BW uses the Key Date as a special parameter to read the relevant attribute records that correspond to the characteristic.

You may have designed queries in the past that use attributes but have never specified a Key Date and found that sometimes the correct information is displayed and other times it isn’t. This is because by default the query will use the current system date to find the attribute information if no Key Date is specified for the query, resulting in only the most recent attribute data being displayed.

To ensure your reports display the correct attribute information all the time, make sure you either manually specify the Key Date in your query design or use a variable to determine the date at run-time. This can be done by selecting the Query Property screen from within the Query Designer.

My preferred method is to always display a parameter that the user must fill when the report is executed.

Technorati Tags: BW, BEx, key date, star schema, 0EMPLOYEE, BEx Analyzer, Time Dependant, Query Designer

Transfer rules and Update rules are the main areas within a BW system where the BW professional can transform your transaction system data into the required format for what ever InfoObject / InfoProvider is being updated.

Quite often when ABAP is used as part of your transfer and update rules there is a need to debug the code to ensure you don’t introduce any problems into your production environment.

When data is loaded into a BW system a background process is used. This means you can not run the data load in debug mode as you would most other programs. What you need to do is run the load program in Simulation mode.

The first step to take is set a breakpoint in you coded Update/Transfer rule. This can be achieved by:

* Setting a breakpoint within the generated program once the rule has been activated.
* Locate the code manually during your debug session and setting a break-point.
* Hard code a breakpoint in the rule as shown below:

• if sy-uname = ‘MYUSERID’.
  • break.
• endif.

My preferred method is the first option. I try not to hard code breakpoints (second option) as I sometimes forget to remove the code once I have finished debugging and the third option takes I find take too long.

To set a break-point in the activated code you have to find the activated program and then search for your code that you want to debug. Follow these 4 steps (Update Rule):

1. View the code for the Update Rule.
   
2. Locate the code that you want to debug and remember some content that will identify the code that you need to debug.
   
3. Come back out of the code and activate your Update rule. Once activated, select “Display Activated program from the Extra’s menu.
   
4. Search for the code that you located in step 2 and set the break-point.
   

Once your beak-point has been set you then need to run the update/transfer rule in simulation mode. To achieve this you must have already loaded data into the PSA. Follow these steps:

1. Navigate to the PSA under the Monitoring section with transaction RSA1. Find the PSA Request that you want to debug and select the ‘Simulate/Cancel update’ option from the right mouse click menu.
   
2. The screen will then show the details tab of the InfoPackage Monitor for the specific request. Right-click on a selected Data-Packet and select Simulate.
   
3. Select the appropriate ‘Activate Debugging’ option for your rule (Update/transfer). Select the radio button to determine the number of records that you want to process in you debugging session and select the ‘Perform Simulation’ button.
   
4. Select the Data Packets, Number of Records and then the continue button.
   
5. The update process will start and your screen will show the debug window. Press ‘F8’ to run through your program. It will then stop at the breakpoint you set earlier.
   

Technorati Tags: Transfer rules, Update rules, BW, InfoObject, InfoProvider, ABAP, PSA, RSA1, InfoPackage, Data-Packet

Listen to a PODCAST of the Blog:
Teklink-Boom & Bust-by Heather (PC Voice)

You know when you look at some of the aquisitions going on today with Internet-based companies. You might be inclined to think that the 1990s dot-com boom era is making a return.

Think about it, eBay bought Skype for $2.6billion. News Corporation, Rupert Murdoch’s company bought MySpace for $580million, and now the word is Yahoo are in talks to buy Facebook for a figure getting close to $1billion.

Now back in the 1990s, it seemed venture capital firms were begging to throw millions of dollars into investments with online-based companies that have next to no track record or decent business plan and we all know what happened then. Something stinky hit the fan.

This time round though, things look a little different. The companies that seem to be being bought, at the least have track records of attracting millions of visitors to their site on a daily basis. For instance, take a look at YouTube. They are quoted as receiving 100 million video downloads a day, with over 70 million viewers a month. It’s pretty much a given that the Google Video venture was a flop, hence their reason for buying YouTube. This has resulted in Google increasing its amount of video streams by at least a factor of 10.

So it seems this time round, internet companies like YouTube are being bought, not just for what they might promise, but also to allow companies access to online communities that are growing faster than ever before.

Most of the companies being bought, also have decent business models and are making money on the Internet by online advertising. And for me this appears to be the key in this round of purchases. It’s all about advertising. Advertising equals big bucks. Look at the figures. I heard Google sold about $10billion worth of advertising last year and returned a profit of over $2billion. Now that’s not to be sniffed at.

I was reading through some BBC technology news where they reported that Google’s YouTube deal is based on the idea that Google will make money from the site’s Content by sharing advertising revenues with commercial video makers such as music companies and amateurs that put their home movies up online. This basically sums up my thoughts exactly.

One of the Internet purchases that has puzzled me though is eBay buying Skype. Skype is a great service, free telephone calls and all, but how can eBay make money out of it? Are they going to stick small banner ads for online auctions on the Skype application? Surely they can’t be thinking about adding quick five second audio advertisements at the start and end of each call.

If you want to buy something from eBay, it would be nice to be able to talk to the seller and asked questions about the product before making a purchase but is it really going to work? The seller would have to be online at the same time as you are making your bid, if not, you would after leave a voicemail and wait for a reply, which is more or less exactly the same process you go through when you email the seller with your question. I guess with this deal will just have to wait and see what happens

A big difference this time round with these Internet company purchases, is that the companies actually making the purchases are Internet or media companies themselves and have substantial cash and revenues. Take Google for instance, who is reported to be sat on almost $10billion in cash. These guys are looking at long-term investments, and not just a quick book.

This doesn’t mean these Internet purchases are all good deals though. Time and time again, we get reports that companies have overpaid in their attempt to gain new customers. The Internet is an unpredictable place, and no doubt some sites will lose their attractiveness to viewers, resulting in costly mistakes.

I have to say though, I am hoping most of these acquisitions work out for the best, and help the Internet emerge from its teenage years, turning it into a young adult that all companies will be happy to do business in.

PS. If Google, Yahoo or Rupert Murdoch (i’m not fussy which), have lots of money to throw around, I’m sure I could dream up some barmy Internet community site and let you throw some money my way. Hah, I can dream.

Cheers,
Pete,

www.TekLink.co.uk

Technorati Tags: PODCAST, Read by Heather, dot-com boom, eBay, Skype, News Corporation, Rupert Murdoch, MySpace, Yahoo, Facebook, investments, business plan, Google Video, online advertising, BBC, free telephone calls, online auctions, www.teklink.co.uk, teklink, pete moxon

GNot has updated his Add-Meta-Tags plugin for WordPress just one day after I installed his previous version and sugested an improvement.

Good going GNot.

As much as I would like to think GNot did the change on my request , I think not. This enhancement was at the top of the To-Do list. I guess the timing was just a coinsident.

Well done, all the same.

I’m also happy to see there are 2 variables included in the plugin file that let you me over write the generated meta-tag entries to what ever I choose.

Cheers,
Pete

www.teklink.co.uk

Technorati Tags: GNot, Add-Meta-Tags, WordPress, meta-tag

I’ve been hunting around for a good plugin that will add Meta-Tag data to my blog posts. I tried a few different ones out but had little niggles with each one until I came across ‘Add-Meta-Tags‘ WordPress Plugin by GNot.

The plugin GNot has written does exactly what it says on the tin.

In summary it will generate the Meta-Tag Keywords and Description for each blog post automatically. For the description it uses an excerpt from your blog page and the Keywords are genrated from your WordPress Categories.

The only thing missing is for the meta-tag info to be auto-generated for the blog index page. This is noted on the To-Do list but for know – It works a treat. Nice job.

Checkout GNot’s page at GNot for the latest version of the plugin.

Technorati Tags: Meta-Tag, Add-Meta-Tags

When users run SAP BW queries and use the F4 key to select values for selections and parameters there is the possibility they will get to see sensitive information they are not authorised to view e.g Payroll information (Pay Grade) which could be part of the Employee master data Infoobject.

To alter the settings that determine the attributes that are displayed to the user, g o to InfoObject Maintenance screen through transaction RSA1 or RSD1 and select the Characteristic that contains the sensitive data as an Attribute.

Select the Attributes tab and there you will see a column called Order for F4 Help. This column determines the sequence in which the attributes are displayed in the F4 help dialog. To stop the sensitive attribute in question from being displayed in the F4 help dialog enter a value of zero in this field.

Technorati Tags: SAP, BW, InfoObject, RSA1, RSD1, Characteristic, Attribute, Order for F4 Help, Business Intelligence, SAP BI

SAP HR users have difficulty understanding how to use the Action Type and Action Reason dimensions on the employee and personnel action Infocube.

The confusion exists when a user tries to restrict the BW query on Action Reason without first choosing the Action Type. When this happened a pop-up window display all the entries of Action Reason independent of any Action Type. Very confusing.

The rule of thumb is ‘Action Reason’ is a sub-division of ‘Action Type’. Therefore, users should always select the Action Type before selection the Action Reason. Doing this will ensure only the relevant Action Reasons are displayed in the pop-up selection window.

Technorati Tags: SAP HR, dimensions, Infocube, BW, Business Intelligence, SAP BI

It is sometimes unclear how BW handles the formatting of dates. Here is a quick overview:

For BEx reporting, SAP BW uses the settings stored on the users local machine. On your desktop click Start -> Settings -> Control Panel -> Regional options to view the settings local to your PC.

Some date related reporting objects get their settings elsewhere.

0FISCPER , 0CALWEEK and 0CALMONTH take their format setting from the SAP BW User Profile default settings (transaction SU01).
0CALDAY though takes it’s setting from the PC Regional Options as shown above.
The format for year is always yyyy(4 digits).

Technorati Tags: BEx, SAP BW, 0FISCPER, 0CALWEEK, 0CALMONTH, SU01, 0CALDAY, Business Intelligence, SAP BI

To resolve problems and answer BW queries, it is often necessary to compare R/3 data with the data reported through BW. Just identifying the Datasource for an extractor can sometimes be a challenge in itself.
For Master Data, Text and Hierarchies take a look at table ‘RODCHABAS‘. This table contains entries of InfoObject/Basic Characteristics together with various entries detailing the source of the object, e.g. Function Module (field ‘MSDFNAME’), View (field ’MSDVIEWNM’ for Master Data objects based on a view). The table also contains indicators such as Object is Time Dependant and Object has Attributes.
If it is a Transaction Datasource you are best investigating table ‘ROIS’. This table lists all the function modules used to generate the data. Select the function module you require and enter the name into transaction SE37.

Technorati Tags: BW, R/3, Datasource, extractor, Master Data, Hierarchies, RODCHABAS, InfoObject, Function Module, Master Data, Time Dependant, Attributes, Transaction Datasource, ROIS, SE37, Business Intelligence, SAP BI

When loading data from R/3, data record errors often occur due to Overlapping Time Period. The reason for this type of error is BW has tried to load multiple records that contain the same key and overlapping validity periods (start date/end date). The records are then rejected during the update.
You can sometimes get around this error by adding additional selection entries to the Infopackagethat load the data. Even though this workaround may work, it should be as a quick fix.
Another, more permanent option is to tailor your data extractor to exclude the invalid data records via ABAP. Even though this may well be sufficient to allow the data load to work, there will still be incorrect data in the source system that needs correcting.
Correcting the invalid data in the source system is the next step but you shouldn’t stop there. You need to investigate how this situation occurred in the first place. Quite often this situation appears in newly installed systems and can be classed as early teething problems with the R/3 configuration and may never occur again. If however users continue to create invalid records, the issue needs to be resolved permanently, e.g through revisiting the configuration and adding thorough data validation.

Technorati Tags: R/3, Overlapping Time Period, BW, Infopackage, ABAP, source system, Business Intelligence, SAP BI

I visit many SAP BI/BW customers in my job and I often notice that the BI//BW team quite often fail to have their DataSource Replication transported through their SAP landscape. When asked why they do not include it in their transports, the normal answer is a simple ‘we don’t know what object to include‘.

With this in mind I thought I’d put together a quick ‘How To‘ document to help out.

• How To…

When you have made a change to a DataSource in R/3 you need to ensure the DataSource is replicated within BW.

To ensure the DataSource is replicated in all target systems when transporting from your development client, ensure you add the DataSource/FileSource (ISFS) object to your transport via the Transport Connection facility within RSA1.

See the screen shot below showing the Employee Object as an example:

Once you have added the DataSource/FileSource object to your transport, the transport should contain the DataSource Replicaobject for you DataSource as shown below:

It’s easy when you know how.

Technorati Tags: BI, BW, DataSource, Replication, SAP landscape, R/3, FileSource, ISFS, Transport Connection, RSA1, DataSource Replica

Check out this YouTube video showing Jeff Hann demonstrating a new way of interfacing with computers. It looks really cool but I don’t see it being the main way we interact with computers in the future. Surely voice recognition is where it is at.

I must admit though, this would be a great way for exploring web and Google Earth.

Minority Report Style Technology

Technorati Tags: YouTube, Jeff Hann, voice recognition, Google Earth, Minority Report

A Step-by-Step guide

Have you every wanted to perform calculations using dates defined as characteristics but have never worked out how it can be done? Replacements Path Variables are the key.

Using Replacement Path in your Text and Formula variables, replace values held in a characteristic. The Characteristic value variables are replaced by the results of a query at run-time.

The steps detailed below show a technique to enable a BEx query user to determine the number of days between two dates.

Scenario:
The group HR administrator wants a detailed line item report that lists all employee absences in a given period. The report is to show the employee number, the absence start date, together with the end date of the absence and show the number of calendar days the employee was absent.

The good thing about using this technique is that no redesign work is needed from your technical BW team, no ABAP is involved and best of all, it quick and easy.

Solution:
For this example I created an ODS object called Absence that holds basic employee information along with individual absence record data.

Follow the steps below:

1. Open the BEx Query Designer and create a new report on your chosen InfoProvider.
2. Drag the Employee, Valid from and Valid to characteristics into the Rows section of the screen.
   If needed, apply data selection restrictions to the characteristics as shown in Figure_1.
3. Right click on the Key Figures structure and select New Formula (Figure 1).
   Figure_1
4. In the new formula window right click on Formula Variable and choose New Variable ( Figure_2).
   Figure_2
   
5. The Variables Wizard will launch and will require you to specify the variable details.
   ( Click the NEXT button if the Introduction screen appears )
6. Enter the variable details on the General Information as shown in Figure 3 .
   Enter the Variable Name , Description and select Replacement Path in the Processing by field.
   Click the Next Button.
   Figure_3
   
7. In the Characteristic screen (Figure 4) select the date characteristic that represents the first date to use in the calculation (From Date).
   Click the Next Button.
   Figure_4
8. In the Replacement Path screen (Figure 5) select Key in the Replace Variable with field. Leave all the other options as they are (The offset values will be set automatically).
   Click the Next Button.
   Figure_5
9. In the Currencies and Units screen (Figure 6) select Date as the Dimension ID.
   Click the Next button.
   Figure_6
   
10. The Save Variable screen (Figure 7) displays a summary of the new variable.
    Click on the Finish button to save the variable.
    Figure_7
11. Repeat steps 4 to 11 to create a second variable for the second date to be used in the calculation.
    In the example shown, the characteristic 0DATETO is used to create the variable ABSEND (Absence End Date).
    

Define the Calculation

We can now use our two new replacement variables to define our new calculated key figure that generates the number of absence days for each record.

1. You will now be back at the New Formula screen (Figure 8). Drag and drop the two new variables into the formula section of the screen and insert the subtract sign (“-“) between the two.
2. Give the new formula a description and click the formula syntax check button to ensure the formula is valid.
   Figure_8
3. The new calculated key figure will now show in the columns section of the BEx query designer (Figure 9).
   Figure_9
   
4. Save the query and execute it.

In the example shown the Number of Calendar Days Absent is calculated correctly. See the table of results below.

Technorati Tags: characteristics, Replacements Path, Formula variables, BEx, ABAP, ODS, BEx Query Designer, InfoProvider, Dimension ID, 0DATETO, calculated key figure, Business Intelligence, SAP BI

Click to enlarge.

An old favourite that is good for a chuckle.

Technorati Tags: How Computers Work, computer sketch

I’ve just found a site that allows you to use their royalty free music in you presentations, blog, Podcasts etc…

Check them out. Some of the sounds are average but most are good. I’ve been looking for royalty free tunes that I can use in my up and coming podcast on SAP BW. Fingers crossed I’ll find something at this site I like.

By the way, all they ask in return is that you add their weblink somewhere in your website…just like I have

If anyone knows of any other good royalty free music I could use, feel free to leave me a note in the comments box.

Cheers,

Pete

Partners In Rhyme Your audio resource for multimedia projects with free royalty free music, sound effects, midi files, audio tutorials, audio software and sound advice.

Technorati Tags: royalty free music, free tunes, podcast, SAP BW, free music, Partners In Rhyme

Hi,

Welcome to TekLink. Fingers crossed I’ll be posting all sorts of gibberish about SAP and technology in general.

 

Cheers,

Pete