So you’ve installed a SAP BW on HANA Developer edition and you’re ready to develop in your own name-space. Here’s a short article on getting the hardware key to submit to SAP.
Finding the Hardware Key for SAP BW 7.4 was straight-forward. HANA Studio (HANA DB 1.0) was a little trickier because the default system was not connected.
Here’s how you fix that.
[thrive_leads id=’384′]
Here’s the thing. Your data model is changes.
Your data loaded quickly and your reports were fast at go-live. Now, it’s chugging along and your users may be complaining (just maybe not to you).
Your company may have acquired a new line of business. Or your model was so well received that additional features were requested.
The pristine design that you built some time ago is now showing it’s age and the unbalanced dimensions are slowing down the reports and data loads.
In the e-book Run Faster SAP BW, I covered the ‘Quick & Dirty’ approach to performance optimization for a quick turns and reduced risk. I won’t repeat it here.
Instead, we’ll elaborate on the ‘Rebalance Dimension’ section.
When grouping items together into a dimension, your goal is to make the size of the dimension as small as possible. In other words, the cross-product or possible combinations of the info-objects in the dimension should be small (so that the record count is less than 10% of the fact-table).
Here are some examples:
+ Dimension: Organization
– Info-Object: Company Code
– Info-Object: Sales Organization
– Info-Object: Plant
+ Dimension: Time
– Info-Object: Calendar Year
– Info-Object: Calendar Month
– Info-Object: Calendar Day
+ Dimension: Product
– Info-Object: Material Number
– Info-Object: Batch Number
– Info-Object: Vendor
In the above example, a parent company will have a small list of sales organization groups and plant locations. It makes sense to group those items into one dimension.
The same thing can be said about the time dimension, since each calendar year will have only 12 months and (generally) 365 days. Because the list is relatively small (and finite), the time dimension is a great for date-related items such as calendar day and fiscal period.
If your company procures raw material, then associating the vendor number to the product makes sense. The finished good could also be related to a certain batch. From a business process stand-point, this makes sense so we will continue the relationship between the items in our model.
What does not make sense is to mix objects as shown in the example below:
+ Dimension: Bad-DIM
– Info-Object: Calendar Day
– Info-Object: Material Number
If the above two fields were keys in a table, then you can imagine the possible day-to-material combinations. On a single day, a large manufacturing company could produce thousands of different products.
A bad design like this could lead to very long run-times during the reporting phase. A simple query should need to scan a large table for calendar-material combinations instead of reading a much smaller table for calendar day and another smaller table for material.
Two quick reads on small tables (in parallel) is much faster than one long read on a large table (in sequence).
So give that a try.
1. Create a copy of an infocube
2. Rebalance the dimensions in the new info cube
3. Load the new infocube
4. Update the statistics of the new infocube
5. Run the program SAP_INFOCUBE_DESIGNS
6. Confirm the dimension size has decreased
Let me know if that helped.
Until next time,
Hau Ngo
Founder + Principle Consultant
If you found this useful, get the free e-book and to catch up with the previous lessons that elaborate on fine-tuning the critical parts of your design.
[thrive_leads id=’377′]
SAP has been refreshingly open in recent years with free online training courses (https://open.sap.com).
To further your understanding of the concepts, SAP has provided trial versions of their software (https://cal.sap.com). They have also streamlined the installation process by using Amazon Web Services. It’s almost as simple as downloading an app from the iTunes or Google App Store.
Almost.
In my first experience with the SAP Marketplace and AWS, I ran into 3 undocumented issues that had me stumped for days. I’ve outlined the issues and resolutions below in hopes that this will save you the headache.
I’ll briefly go over the three issues & resolutions in the paragraphs below. If you need additional clarification, then check out the video at the end of this article where I go into further detail.
In order to install the trial, you need a AWS user-id with access to the credentials as per http://aws.amazon.com/developers/access-keys/
Just remember to assign your user-id to the new groups.
If you happen to see the a “pending verification” status during the installation of your SAW BW on HANA Trial, then you need to take three additional steps.
Having received this message in a prior attempt, I waited 20 hours only to find that the installation was stuck. Just delete the instance after you contact Amazon and install it again.
If you’ve gotten this far, then you should see a green square next to your instance in SAP CAL, indicating your appliance has been installed, active, and ready for use. Congratulations!
However, you cannot tunnel into your server using Remote Desktop.
To resolve this issue, delete your instance and recreate it (again) with the following setting
There seems to be region specific issues with the SAP Trial packages. I tried a few combinations during the installation and the above setting worked.
As an added bonus, there is an Wonderful Guide that provides plenty of exercises (and step-by-step instructions) for the new functionality in SAP BW 7.4 on HANA and SAP BusinessObjects 4.1 on HANA. I highly recommend going through the exercises to become more familiar with the new concepts.
The following video explains the 3 issues in further detail. Check it out and let me know if the video helps?
http://scn.sap.com/docs/DOC-54564
A how-to for extracting activity data from Fitbit and analyzing it with an enterprise-grade visualization tool
For one who dislike wearing watches or accessories, I have surprisingly worn the new Fitbit Charge HR (heart rate + activity wristband) every day since it was released for a solid two months.
The idea of effortlessly collecting data to gauge my activity is persuasive and the new clasp is solid, holding up after an hour of kickboxing.
I also love the companion app for the iPhone. The UI is gorgeous and you can quickly view the various metrics for today and previous days. Here are some screens.
[envira-gallery slug=”fitbit-mobile-ui-analytics”]
After using it for a few weeks, however, I found a limitation in the default UI.
For example, metrics such as calories burned or total steps can be viewed one day at a time. You need to toggle to another day to see a different set of metrics. There is no easy way to cross reference metrics and draw a correlation between an activity and calories burned.
Since I needed simple data exploration, the decision to use SAP Lumira was easy. But how do I export the Fitbit data? Luckily, there was a published recipe on IFTTT to automatically import your Fitbit data into Google Sheets.
That’s it! Just follow the prompts at step 3 and you’re set. Each morning, the IFTTT app will alert you to a new entry, where new record will be appended to your Google Sheet.
One manual step is a description of the metrics. You will need to manually insert a header line into your Google Sheet to describe the metrics. This can be easily found in the IFTTT recipe. Here’s a copy of my sheet for your convenience.
After a few days of collecting the Fitbit data, you should have a solid dataset to create your own visualizations. Here’s mine (PDF link).
With a limited set of data for one week, I can already draw correlations between the increase in one activity such as number of steps taken to the increase of calories burned for a given day. Similarly, fewer calories are burned when I am sedentary. Of course, you don’t need a high-tech gadget and enterprise level analytics to draw that conclusion.
Where this information becomes interesting is in the application. I have a goal of burning 2,900 calories per day but this is a result and not an action that I take directly. Based on the above charts, however, I can see a correlation between calories burned and the number of steps taken, which is an action that I can directly influence.
As a result, I can set this metric (10,000 steps) as a daily goal, which triggers an alert vibration when accomplished. When I feel the a vibration from the Fitbit Charge HR, I know that my personal daily goal of burning approximately 2,900 calories has been accomplished.
As of the writing of this article, the heart rate metrics are not available for export but the folks at Fitbit will make that data available soon.
For intraday information (minute-by-minute data), you will need to request partner access to their API. Here’s a great article on setting that up.
A client asked for a custom business intelligence report to track the aging of her products in quarterly buckets (1-3 months, 4-6 months, etc.). She could not afford to assume that her quantity of stock on hand is actively moving due to one recent goods receipt, as suggested by SAP Standard Content. She needed additional insight into her data.
“Can we build a more accurate inventory aging report? The one provided by SAP ages the entire stock quantity by the latest receipt date. This does not give us a valid aging of the stock that we have on hand.”
While the standard content for SAP Business Intelligence is immensely useful for quickly implementing common solutions and getting critical reports into the hands of business users, the out-of-the-box material do not address specific issues due to the different nature of each industry.
Imagine a grocery store manager who needs to move his inventory of milk and we can begin to see the urgency.
Assuming the manager receives the bulk of his delivery on the first of the month, he will need to to quickly start moving his product by the 15th of the month before his product spoils. This can be accomplished with a sale or promotion. A delivery of additional milk products at the end of the month should not imply that his remaining stock is fresh.
Luckily, most of my clients do not stock perishable goods. Still, having slow or non moving stock hinders a company’s ability to operate since the capital is locked-up as inventory.
Figure 1. The out-of-the-box inventory aging report, which applies the latest goods receipt to the entire stock
First-In First-Out (FIFO) is easily understood on a conceptual level. You ship the material in the order in which you received it.
However, this approach is difficult to implement at the programming level because of the book-keeping overhead. There isn’t a direct relationship between a goods receipt and a goods issue in the logistics module. You will need to keep an account of the movements manually.
Tracking the stock balance while applying the latest goods issue to the earliest goods receipt required an large amount of coding and testing, which we could not afford in our development timeline.
To work around this constraint, I developed a simpler solution. The approach hinged on the assumption that the earliest goods receipts will be consumed by the earliest goods issue. As such, they do not need to be accounted for.
Because I had already helped this client to implement the standard non-cumulative inventory solution to report current stock balances, we already knew the ending balance quantities and values for a given set of materials on a given date.
With the ending balance quantities identified, we simply worked backwards and counted only the goods receipts that contributed to the current stock.
Figure 2. A custom inventory aging report with various buckets for segmentation.
The client loved the simpler approach and I was able to deliver the solution in a few days, well within the constraints of the project timeline. If there were an opportunity to revisit this solution, I would like to further increase the accuracy of the valuation to account for variations in the monthly material valuations (used at other clients). More to come in a future post.
What do you think of this approach? Would you like more detail on this solution? Does it address an issue that you’re currently facing? Send me an email or let me know in the comments below.
P.S. Looking to implement your own custom solution? Get my pre-made guides where I map out the information that you need.
[thrive_leads id=’346′]
Recently, a client has asked for a custom datasource and transformation for loading unit-testing data into a quality assurance environment.
I ran into some common problems so I created a reference document to speed up the process. If you find this useful, please leave a comment and let me know if it helped.
Summerlin Analytics SAP BW Flat-File Load
Did you like the how-to guide and want to learn more? Check out my ebook with practical tips for optimizing your data models.
