Skip to Content
|

Category Archives: Uncategorized

Folks,

I’ve already written before about the lack of motivation to blog on SAP Community (SC). It’s been over a year but I don’t see any improvements. The blogging guidelines have been quietly swiped under the carpet. It’s not clear yet how this will help the quality of content or engagement. Anecdotal evidence suggests that the numbers of views, likes, and comments dipped even further.

But we can write and read blogs on many other websites and it’s not what this is about. So far SC is still the best place where one can find a good answer to many SAP questions. There has been no serious competition in this area and I imagine it remains attractive to the SAP customers.

However, since 2016, SAP managed to fail in attracting active new SC members and simultaneously succeeded in alienating the veterans. How did that happen?

You say SAP community, I say SAP Community

The renaming of SCN to SAP Community in 2016 was one of the poorly thought through decisions. And constantly having to type “SAP Community” instead of shorter SCN is not its only ill effect.

Now when anyone mentions “SAP community” it’s not even clear what are we talking about. Is it “SAP community” (as a group of people who work with SAP products) or “SAP Community” (as in the website and successor of SCN)?

There is a part of SAP community that does not actively participate on SAP Community (although most likely consumes the content occasionally). There is a part of SAP Community that does not participate in in-person community events, such as SIT, conferences or user group meetings. There is a part of SAP community where “cool s*t happens” and there is a part where s*t happens without being cool.

But when we just proclaim everything and anything “SAP community” it seems to make it easier to not continue including the less “cool” parts of whatever “SAP community” is. “What do you mean SAP does not support SAP community? Look at all the spending and all the successful SITs and CodeJams!” – “Um, not THAT SAP Community…” – “Huh? It says ‘SAP community’ right here, sooo…”.

SAP also made it very clear that they want to aggressively pursue non-SAP developers and, ultimately, include them in “SAP community”. The cool kids are probably not going to want to hang out on the non-cool SAP Community. So where does this leave us, exactly?

The Platform is Broken

The half-baked re-design of 2016, rushed into production just so that a grand announcement could be made at TechEd, has not been an improvement for the SCN users. 2 years later, SC remains a tool equally frustrating for the veterans and newbies. (This is where full inclusion has already been accomplished, ironically.)

As a company that heavily invests in their own platform, SAP of all organizations should know that you can’t build anything good on a broken platform. Running SC on a broken platform and expecting a different result is a definition of insanity.

Change of The Guard. Or Not

Do you know that Jürgen L, perennial Top Contributor, one of the tenured and most respected SCN members, already left the building? It’s easy to see the result of his absence:

I ran the statistics from the 27 questions in MM tag right before the Jürgen’s departure.

For comparison, similar 27 questions from SAP HANA tag (mind you, here “SAP employees” include well known top contributors, just like Jürgen):

It’s not hard to see that in many tags the same 2-3 people answer majority of questions. This just shows how much we rely on our most dedicated members to offer expertise to share.

One might say, oh boo-hoo, good riddance, time for the old fogies to step down and make space for the new generation. Fair enough. But where is this new generation?

The Stack has Overfloweth

SC is not the only online community going through the rough times. StackOverflow (SO) has been recently undergoing changes to address its reputation as an “elitist” place where no one dares to ask a question anymore.

One of the issues common between two communities is lack of guidance and support for the new members. SO makes use of technology to resolve that. SCN’s solution? Simply declare that from now on we will be “inclusive”.

While searching for a different perspective on the happenings at SO, I stumbled upon this old Reddit post. It is an illuminating read and if you’re truly interested in the community matters I’d recommend reading the comments all the way through.

The first comments support the image of SO as a cliquey website where one cannot post a question without being shut down and dragged through the mud. I will not deny that there are cases when on SC we (myself included) could’ve acted nicer too. But the later comments paint the eerily familiar picture of the active contributors burned out from repeatedly answering the same homework questions.

The Social Contract

StackOverflow code of conduct establishes a clear social contract. If you’re here to ask then the answer to your question will depend on your own effort and someone’s kindness. If you’re here to help then be patient. As much as I love to criticize anything and everything, I’d have hard time disagreeing with this.

SCN used to be a place by professionals and for professionals. But over the years, that social contract, as we say in the US South, “got gone”.

The inclusion that is not supported by platform features and human actions is just a bunch of baloney. And encouraging the drive-by, “dear gurus, here is my client requirement” questions for the sake of generating the content mass is not the inclusion. It is an enablement mechanism for the shady consulting companies.

Those who expect to be treated like professionals need to behave as such.

So what now?

Without a clear definition of (and adherence to) the social contract and a functional platform SAP Community will continue to decline. SAP community, of course, won’t cease to exist. But where will it go? That is the question we need to start asking ourselves. The ball has been in the SAP’s court for so long on this that the game might as well be considered forfeited.

Start with a Solid Process Baseline

Part 3 of a 5-part Series

Enterprise business processes are designed to support operations. But it’s not always easy to understand how well these processes are really performing or where they need improvement. To identify process weaknesses, compliance deviations, and opportunities for optimization, business process owners need complete transparency into how processes are executed.

To measure the efficiency and conformance of business processes, they need to be able to objectively assess how well processes execute. By using data-based process discovery, organizations can identify all process variants and bottlenecks that may increase costs or lead to compliance problems.

Visualizing and understanding the operative processes running within enterprise applications, and distilling large volumes of data about business process execution into intuitive graphical representations, can provide a clear picture of how tasks are being performed, how well they run, and where they consume the most manual effort.

Organizations can also understand how things work and where to remediate inefficient, cost-intensive, and noncompliant processes. By enabling complete transparency into as-is processes, companies can become more compliant, increase efficiency and enable data-based continuous process improvement.

At SAPPHIRENow, SAP Market Influencer Eric Kavanagh recently sat down with Bastian Nominacher, co-CEO of Celonis, to discuss the importance of understanding how and where data is flowing when undertaking a compliance project, or a transformation initiative.

Bastian’s Recommendation

“You need to have a starting point for any transformation or compliance project. You need to have a solid baseline and a way of verifying where you stand, and them optimize based on this.”

Watch the Interview with Celonis

To see the third of the 5-part discussion, watch the video.

Did you miss Part 1?

To see the first of the 5-part discussion, click here.

Did you miss Part 2?

To see the second of the 5-part discussion, click here.

Learn more

Learn more about process mining.

 

Thanks to Andrew Lunde,Thomas Jung, Tom Slee for all the valuable inputs.

This blog will cover 4 steps in end to end development and deployment of the app to cf.

  1. Create the Hana database using the Hana Service (DBaaS) on cloud foundry (CF)

  2. Configure Web IDE Full stack to work with cloud foundry.

  3. Deploy the App on CF.

  4. Debugging using webide fullstack.

 

Step 1: Configure Hana service on CF.

 

Goto space -> Service marketplace -> SAP Hana Service.

 

 

Click on the different plans based on your entitlements.

Set the password and other parameters like size .(eg min size is 32 GB and can be increased in 16GB block increments)

Step 2: Configure Web IDE Full stack to work with cloud foundry.

Web IDE fullstack is available on  neo but not on CF.

So , we need to configure the webideFullStack from neo to work with CF. Basically, we need to deploy a builder app that would be used for building/deploying apps built in webidefullstack.

Goto neo subaccount -> Services -> Web IDE Full stack

Click on Configure service and then provide necessary authorizations to your user (say P12345).

 

Now click on goto service which opens the webide full-stack.

In web ide fullstack, Goto Settings -> cloud foundry ->

Select your CF endpoint. You will be asked for CF login emailed / password. After you enter the login credentials, select the Org & Space to deploy the builder App.

 

Finally click on the Install builder and save it.

 

Also enable hana and Node dev features in Web IDE full stack as shown below.

 

Now if you goto your space in the CF environment, you should be able to see the builder app(webide-builde-sapwebide-ID) under your applications.

 

Step 3: Deploying a nodejs application to CF.

 

Goto webide fullstack and create a new project from template.

 

 

 

Now create a new Hana Database module in the MTA.(Multitarget application)

 

 

 

 

Make sure to select the correct version. Based on the info in the Hana service cockpit.

 

i.e goto the Hana cockpit to find the version info in cloud foundry. CF->space->services -> service market place -> Hana service -> instance

Also make a note of the Database ID highlighted below. Hana service dashboard*****

Now create a new node Js module in webide full stack.

Enable the XSJS support.

 

 

Now click on mta.yaml and use the MTA editor.

Click on the nodejs module and select requires(+). And select the db module & save.

Now edit the mta.yaml file to configure the Hana DBaaS settings. Refer Hana service dashboard*****

above for the database ID. Use space instead of tabs to avoid indentation errors.

First build the DB module and then the nodejs module.i.e right click -> build module.

 

You can edit the Index.xsjs in the lib folder of the node module to establish a db connection and do a select from the dummy table.

 

$.response.contentType = “text/html”;

var output = “Welcome <br><br>”;

var connectn = $.db.getConnection(); //Open a database conn

//PrepareSQL statement on the sys table “DUMMY”

var pstatement = connectn.prepareStatement(“select * from DUMMY”);

var result = pstatement.executeQuery();

//Check the query result

if (!result.next()) {

$.response.setBody(“Failed to retrieve info!!”); // Return error

$.response.status = $.net.http.INTERNAL_SERVER_ERROR;

} else {

output = output + “Query execution result:  ” + result.getString(1);

}

result.close();

pstatement.close();

connectn.close(); //Close the  connections

$.response.setBody(output);

 

Now , if you click on node module & Run -> Run as node application, you will see a an application deployed and running on CF under applications. This might take couple of minutes.

 

 

 

4.Debugging using WebIDE fullstack.

Start the application.

Click on the debug icon on the right as shown below.

Click on the connect icon to attach the debugger to the application.

Set some breakpoint at the beginning of index.xsjs.

Rerun the application url,

 

Now you will be able to see the control coming to the breakpoint in index.xsjs.

 

 

 

As we become increasingly surrounded by innovations like artificial intelligence, machine learning, and chatbots, business leaders are ironically trying to find ways to make their companies more “human”. One of the key mandates of today’s CHRO is to create personalized, exceptional experiences for every member of the workforce, ensuring engagement and productivity across functional boundaries. But what we’re increasingly learning from companies that lead in digital transformation is that the employee experience can be “humanized” not in spite of, but because of technology and data.

Tune in live Wednesday, August 15th or access the link on demand to hear the experts speak on how employee experience has been shaped by technological innovations. Hear from:

  • Penny Stoker, Global HR Services Leader at EY
  • Thomas Damgaard, Managing Director in Accenture’s Management Consulting Practice
  • Gabby Burlacu, Solution Manager for SAP SuccessFactors Web and Mobile Experiences

and gain insight into how you can leverage trends in employee experience, artificial intelligence, and big data in HR to create a competitive, unparalleled experience for your people.

 

This show is part of the SAP Radio series Changing the Game with HR. Today’s dramatic business and workforce changes make it urgent to think differently about HR. Instead of being just “the system of record” or “engagement”, HR needs to become an agile platform for everything in your organization to come together to transform the work experience. How can you develop your key relationships across the business as you transform HR into a powerful force for business breakthroughs? Tune in on the VoiceAmerica Business Channel for Changing the Game with HR, presented by SAP, and check out on-demand recordings of past shows.

In this blog, you can find the latest collection of new capabilities in SAP S/4HANA Cloud in the area of Plan to Product / Manufacturing with the focus on the ERP End-to-End processes of Make to Order and Make to Stock.

Here you find the core processes addressing the main production environments Make to Stock and Make to Order:

Make-to-Stock Production – Discrete Manufacturing (‏BJ5‏)

Make-to-Stock Production with Variant Configuration (‏21D‏)

Make-to-Order Production – Finished Goods Sales and Final Assembly (‏BJE‏)

Make-to-Order Production with Variant Configuration (‏1YT‏)

Innovation topic here is:

Demand-Driven Replenishment Planning and Execution (‏2QI‏)

 

What’s new in SAP S/4HANA Cloud for Manufacturing & Supply Chain?

Check our quartely product management updates:

SAP S/4HANA Cloud 1808 Release:
https://blogs.sap.com/2018/08/08/intelligent-erp-update-sap-s4hana-cloud-1808-for-manufacturing/

SAP S/4HANA Cloud 1805 Release:
https://blogs.sap.com/2018/05/09/product-management-update-sap-s4hana-cloud-1805-for-manufacturing/

SAP S/4HANA Cloud 1802 Release:
https://blogs.sap.com/2018/02/14/product-management-update-sap-s4hana-cloud-1802-for-manufacturing/

 

Where to find out more details on the product?

Follow us via @SAP and #S4HANA, or myself via @GerhardWelker.

Neues Feature „Vorschaufenster“

Mit dem SAP Customer Checkout Manager können Sie Ihre eigene hierarchische Baumstruktur aufbauen. Diese Funktion steht unter der Kachel Schnellauswahl unter dem Punkt Kassenkonfiguration zur Verfügung. Mit dem neuesten Update steht nicht nur eine zentrale Nutzung der Schnellauswahl zur Verfügung. Zukünftig werden Änderungen in der Schnellauswahl direkt in einem Vorschaufenster angezeigt. So haben Sie die Möglichkeit weitere Unterkategorien hinzuzufügen und einzelne Menüeinträge in Form und Farbe direkt zu ändern. Die im Vorschaufenster dargestellte Schnellauswahl, wird später an den Kassen Oberflächen genau so dargestellt. Folgende Änderungen sind durch den Benutzer steuerbar:

  • Beschriftung der Schaltflächen
  • Hinterlegen von Bildern
  • Dimensionen anpassen (Höhe u. Breite)
  • Koordinaten bestimmen (Y-Achse; X-Achse)
  • Farben anpassen (Schriftfarbe; Hintergrundfarbe)

In frühersten Versionen konnte die Schnellauswahl nur direkt an den Kassen (lokal) angepasst werden. Mit der Funktion Export/Import konnte die Schnellauswahl-Konfiguration auf weitere Kassen kopiert werden. Das hat die Konfiguration und Einrichtung wesentlich umfangreicher gestaltet. Anbei mein Quick-Guide für die richtige Nutzung der zentralen Schnellauswahl

  • Einsatz des SAP Customer Checkout Managers
  • Aktivierung der zentralen Schnellauswahl
  • Aufbau der Quick Selektion Maske
  • Verknüpfung mit dem Artikelstamm
  • Zuweisung an POS-Instanzen

Verschieben

Sie können Ihre Schnellauswahlkonfiguration auf der Registerkarte „Schnellauswahl“ verschieben. Mit dem Verschieben der Baumstruktur lassen sich die einzelnen Unterkategorien je nach Einsatz neu zuordnen. Nutzen Sie dafür das Symbol „Verschieben“ und wählen Sie anschließend die übergeordnete Ebene, zu der Sie die Schnellauswahl verschieben möchten.

Zuordnung

Mit dem SAP Customer Checkout Manager haben Sie die Möglichkeit die erstellte Schnellauswahl einer einzelnen Kasse oder einer ganzen Organisationsgruppe zuzuordnen. Dabei ist zu beachten, dass jeder Organisationseinheit nur einer Schnellauswahl zugeordnet werden kann. Darüber hinaus kann die fertige Schnellauswahl auch einem bestimmten Benutzer (z.B. Mittagsmenü) oder einer zentralen Rolle (z.B. Kassierer Mittagstisch) zugeordnet werden. Beispielsweise in der Gastronomie, kann so den Benutzern zur Mittagszeit eine eigene Schnelltasten Verwaltung zur Verfügung gestellt werden. Im Back-Office haben Sie zudem jederzeit die Möglichkeit, die Speisen und Getränke bereits im Vorfeld frühzeitig anzupassen. Die Auswahl der Schnellauswahl erfolgt über die Kachel Kassenkonfiguration in der Registerkarte Schnellauswahl. Wählen Sie anschließend die entsprechenden Details in der Konfiguration aus.

Neben der Zuordnung verfügt der SAP Customer Checkout Manager noch über weitere Funktionen zur Konfiguration der Schnellauswahl. Das zählen beispielsweise:

  • Kopieren der Schnellauswahl
  • Definition benutzerdefinierter Textbausteine (Sie dienen dazu, den Verkaufspositionen zusätzliche Informationen anzuhängen)
  • Weitere Drucktasten mit Spezialfunktionen zur Schnellauswahl hinzufügen

SAP beliefs customer feedback is so important. Your feedback helps us to

·        Improve the products, services & experience

·        Improvise the ways to measure delighting customers

·        Take tangible actions to make better decisions which enables better customer satisfaction

·        Identify the needs of customer advocacy

 

1.      How does this work with SAP Business ByDesign Product Support?

Customers provide feedback on resolution of single incidents. Customer will get an email notification with the survey link per incident while they perform the complete action on the incident. The below snapshot shows the action complete from the drop down.

Customer receives the email notification with a survey link – Customer Feedback on Incident. Here is a sample snapshot.

The survey will have the following questions

You can only submit the feedback ones. If you try again, the below message appears. Please take the time to provide your feedback and send it back to us

2.      How to interpret the Survey Scale

SAP started measuring very satisfied feedback from our customer via the rating 9 or 10 since we want to continuously evolve our delivery and delight you, always. The survey questions will focus on your overall experience with support & product

3.      What happens when you rate an incident?

Support Management team will review the feedback and take immediate corrective action if required even in conjunction with the customer. If you rate 9 or 10 to questions below, we will make sure that the team(s) and the employee(s) involved know that they have done a great job. If you rate below 5 due to severe dis-satisfaction; our management team will review the incident handling and personally reach out to you via phone call.  

4.      Who can guide you with more information if required?

Please reach out to your Customer Success Manager or Customer Engagement Executive if you have one. If you don’t have, please ask your queries while you interact with cloud product support.

 

Walk through SAP HANA Data Management Suite

Philosophy, why, and the future

Product, value

How customers can adopt

Usual legal disclaimer applies

Source: SAP

Moving to a decision driven culture

Big data term less popular

Better data, quality of decisions

Organizations are less comfortable with the quality of data than in the past, seeing flaws

Source: SAP

White board of data landscapes – an aggregate of diagrams

Fragmentation is more onerous – graph, spatial, sentiment, making the data landscape more complex

Cloud data sources are new, making the landscape more complex

If you don’t know data exists, you can’t make it secure.

Source: SAP

Data is less accessible, more fragmented

Moving towards machine learning, artificial intelligence, requiring high quality to be trained on

What customers want is on the right based on focus groups, studies

Source: SAP

The vision of HANA Data Management suite

Source: SAP

This is not the 1990’s data warehouse is

At run time, know where the data is, run it at run time, may replicate data

Source: SAP

The value is on the left

Source: SAP

Vision introduced at SAPPHIRENOW

Source: SAP

4 key products in the middle

Source: SAP

Product was launched in June

Future- harmonize the user interface between the products, single user experience

Invest in metadata catalog to improve understanding of data

Cloud – easy to provision, elasticity, move workloads, making this a true cloud-based solution

 

Source: SAP

Process to improve data, continuous cycle

Source: SAP

Ways to connect the data, connect remote data sources

Source: SAP

Time series, streaming data, data becomes “intelligent”

Source: SAP

Multiple types of workloads

Source: SAP

Cloud architectures in many different places, deploy to different places

Roadmap to be released soon

Source: SAP

Three adoption patterns

Strategic – data framework

Tactical – insight

Source: SAP

Where to start strategically?

Source: SAP

Where to start tactically?

Source: SAP

Data Governance – lots of discussion surrounding MDG

Source: SAP

Summary of webcast

Q&A

Q: what parts of the data management suite are in GA?
A: all are in GA

Q: What is the biggest value in your opinion

A: Trends changing landscape – artificial intelligence, machine learning

This series of blogs is to share our experience, tips/tweaks on various integrations provided by SAP between HCM systems; on-premise and cloud.

 

Data Replication/Configuration:

First things first, Employee Central (E.C) -> Employee Central Payroll (E.C.P) data replication via PTP – it is fast. Actually, it is super-fast compared to Core Hybrid (E.C -> SAP HCM) and Side-by-Side (SAP HCM -> E.C) replication.

Of course, the corpus of data that is being replicated is less, no O.M, no middlewares, no pack size issues/timeouts and no daisy chained P.I pass thru systems!

The initial replication of 2500+ employees took less than 3 minutes. Pretty Impressive! The configuration/customization is also super easy.

 

Setting connection for Employee Central System – For DC 12 customers only:

Table RFCDISPLAY shows api12.successfactors.eu. Simply set the connection to api012.successfactors.eu to resolve the ping errors.

 

Code Value Mapping (T77SFEC_CVMAPC): 

Code Value Mapping is 1:1. For example, you map event “Termination” in E.C to action Termination in E.C.P. Same goes for event reason and action reason.

 

 

As you see in the screenshot below, retirement “event reason is configured under Termination Action. When you map these events in code value mapping, the desired employment status of „retirees“ is lost via replication. This happens because Termination event usually has “inactive” status in E.C.P.

Enhancement implementation of HRSFEC_B_CE_CHANGE_IT0000 could be used to resolve this issue.

 

Permanent Establishment Key Mapping (T77SFEC_KMAPPEST): 

Location in E.C is mapped to Personnel Area and Subarea in E.C.P. However, a location can be shared between multiple company codes in E.C (see below). Replicating such an employee will return an error as P.S.A to CoCd has 1:1 mapping in E.C.P.

 

 

Creating additional locations in E.C was not feasible for us, but we had to do that in E.C.P. This had to be handled via Enhancement implementation of HRSFEC_B_CE_CHANGE_IT0001.

Be aware that you will have to register and fill BTRTL, SACHP, VDSK1, SBMOD along with WERKS!

 

Global Information/Challenge Infotype:

Global Info or mapping to Challenge infotype is not available out of the box. This information is critical for running payroll.

Enhancement Implementation of HRSFEC_B_CE_PROCESS_EMPLOYEE can be used here. Value mappings (challenge type, Agency etc. could be stored in T77SFEC_CVMAPC for ease of maintenance.

 

Address Information – Distance in km (ENKTM):

For employees with company car information, it is required to capture distance in km from primary residence to payroll purposes in Germany.

The distance information makes the most sense to be configured under the Address information object in E.C (similar to SAP address infotype).

Basic extensibility (V_HRSFEC_EXTMAP) allows you to map to Job Information, Personal Information, and Compensation Information to the Employee Central Payroll infotypes 0001, 0002, 0007, and 0008 only.

In order to realize this configuration, Enhancement implementation of HRSFEC_B_CE_CHANGE_IT0006 can be used.

 

USERID/PERNR Mapping:

Standard delivery of E.C.P generates PERNR using an internal number range. This means that USERID from E.C and PERNR from E.C.P is different.

Having different employee numbers appears to be a non-issue at first. You maintain all data of the employee from E.C. The mashup screens internally map the numbers and Payroll Control Center is used to run the payroll.

As we all know, payroll does not end with Finance posting and bank transfer. There are a lot of post payroll transactions/government communications in every country. Having different employee numbers was an inconvenience to our customer who had to execute different reports.

SAP provides HRSFEC_B_CE_EXT_PERNR_MAP to resolve this issue. This allows you to keep the employee numbers the same. However, please go thru the METHOD comments before implementation. Consider the impact of global employment or intercompany transfers.

 

Intercompany transfers:

The standard code creates new PERNR during Intercompany transfers. Enhance HRSFEC_B_CE_DECIDE_HIRE_REHIRE to control this. However, please go thru the METHOD comments before implementation.

 

Full Transmission Start Date (FTSD) and Impact on Replication:

E.C.P allows you to set an FTSD. This is very important as the entire history of the employees is not relevant for payroll, data for the current calendar year or financial year would do.

For example, if you have set an FTSD of 01.01.2018, all valid employees on or after this date are replicated. You also have employees retired on 31.12.2017 and who are eligible for retiree pay effective 01.01.2018.

Replication of such cases usually errors out due to:

  1. Lack of “job information” data slice in E.C
  2. Employment status issue as mentioned at the beginning of the document
  3. E.C.P refusing to create new PERNR because of retirement action.

A combination of enhancement, „F.C“ changes in T529A and separate „actions“ can resolve this issue.

 

Payroll Control Center (P.C.C)- Detailed Results in Simulation Posting:

SAP’s RDS delivery covers basic payroll activities (simulation, live run, simulation and live payroll posting, Bank document creation) using a neat and clean looking UI.

(Since P.C.C is all ABAP, I am sure you can cover more process via customer enhancements, but is it worth the investment?)

However, the details of the simulation posting are not displayed. You get to know the document has errors but cannot see any details. The fix is easy:

Enhancements:

  1. Create a new run time class for the step simulation posting (Copy of CL_PYC_STT_CRT_SIM_POSTING_RUN)
  1. Enhance Methods FP3_BPC_DET_RT_GET_LIST, GET_RPT_CHAIN_MODE

For the detailed KBA, please contact SAP support.

 

 

 

Post Payroll – Custom Tiles in E.C:

There are a lot of post payroll transactions for Germany, really a lot J. I am sure this is the same case for all countries. P.C.C does not cover these transactions, this means the payroll admin will have to logon to E.C.P via GUI (username/password).

A simple workaround here is to create E.C Tiles. Each folder that you have under Payroll Germany for post payroll was created as a tile. An example below “Sozialversicherung“ which is only available to Payroll Admins.

 

 

 

The individual tiles calls the different SAP transactions. All your transactions in Tiles instead of folders!

 

This opens the respective report from SAP directly, SSO respected. No GUI required, only WebGUI 🙂

 

 

 

Screen Personas

 

Compared to the good-looking Fiorized SF screens, the humble „PC00_M01_RPCBNVD0_O“ displayed via WebGUI looks uninteresting (see screenshot above). This difference in U.I between systems was not well received.

So what can be done? Screen Personas could help to do some leveling. A nice Landing page, for example definitely enhances the UX.

 

The same „PC00_M01_RPCBNVD0_O“, now looks relatively better 🙂

 

 

 

Wagetype Default from Payscale Structure in E.C.

A simple business rule that factors F.T.E with wage type amount from the payscale structure, rounds and assigns to the recurring object is really helpful for your payroll team, especially if you have lots of TARIFF employees.

 

 

Infotype 0007/0008

Depending on if you are using E.C Time Off or a third party time provider, the integration design of IT0007/0008 varies. This is sensitive to payroll calculation, few thoughts:

  1. F.T.E in E.C is relevant for payroll depending on integration design.
  2. Standard Hours vs Individual hours of an employee effects Cap Utilization Level/Employment Percentage in E.C.P depending on your integration design.
  3. Using a custom field in Job Info for W.S.R, you can replicate 0007 and 0008 together. Standard integration neatly handles BSGRD/DIVGV linkage.

 

Concluding Remarks:

Different customers are in different stages of the digital transformation journey. SAP provides multiple options/product lines to support this transition.

Do talk with SAP or your S.I to understand what is best suited for your enterprise. Our customer transitioned from an on-prem legacy HR/payroll system to the Full Cloud model (E.C, E.C.P, PM/GM and Recruiting).

 

Boilerplate 🙂 It is possible that items mentioned in this blog are resolved via fixes or design changes. Refer the standard integration handbook or quarterly release notes from SAP for the latest product details and for design decisions.

 

 

Hello,

I developed useful functionality for SPRO

I detected that is common the necessity of copy IMG path when you navigated of SPRO Tree Customizing .

With the following code you can copy the path very easy

The change is only apply in Developer System and is necessary register one standard object.

  • Step 1: Add customer Button in Status GUI IMG_DISP of Function Group S_IMG_TOOL_1.

Register the object in Marketplace, before the changes apply

  • Create a Custom Function Code “ZCOPY_P”

Add custom Function Code “ZCOPY_P”

  • Add Custom Code Functions in Interface and Key Function

  • Step 2: Insert Implicit Enhancement in INCLUDE: LSHI01F0X – end of FORM  execute_user_command

  • Step 3: Add the following ABAP code and active the implicit enhancement
DATA: lv_nodekey TYPE tv_nodekey ,
      le_item TYPE shi_item,
      le_node TYPE treev_node,
      lv_relatkey TYPE tv_nodekey,
      lv_text(500) TYPE c ,
      lv_relatship_ant TYPE int4,
      lt_path TYPE STANDARD TABLE OF char100,
      le_path TYPE char100,
      lv_choise TYPE  sy-tabix.

CASE fcode.
  WHEN 'ZCOPY_P'.
    CHECK g_tree_data->tree IS NOT INITIAL.
    READ TABLE g_tree_data->nodes INTO le_node WITH KEY node_key =  g_tree_data->node_key  .
    CHECK sy-subrc = 0.
    READ TABLE g_tree_data->items INTO le_item WITH KEY node_key =  g_tree_data->node_key  item_name = 'TEXT'.
    CHECK sy-subrc = 0.
    le_path =  le_item-text.
    APPEND le_path TO lt_path.
    lv_relatkey  = le_node-relatkey.
    lv_relatship_ant = le_node-relatship.
    WHILE lv_relatkey IS NOT INITIAL.

      READ TABLE g_tree_data->nodes INTO le_node WITH KEY node_key =  lv_relatkey .

      READ TABLE g_tree_data->items INTO le_item WITH KEY node_key =  le_node-node_key item_name = 'TEXT' .

      lv_relatkey = le_node-relatkey .

      IF  le_item-text IS NOT INITIAL AND  lv_relatship_ant = '4'.
        CONCATENATE le_item-text '->' INTO le_path.
*        le_path =  le_item-text.
        INSERT le_path INTO lt_path  INDEX 1.
      ENDIF.
      lv_relatship_ant =  le_node-relatship..
    ENDWHILE.

    CALL FUNCTION 'POPUP_WITH_TABLE_DISPLAY'
      EXPORTING
        endpos_col   = 130
        endpos_row   = 10
        startpos_col = 40
        startpos_row = 1
        titletext    = 'Ruta de Configuración'
*      IMPORTING
*        choise       = lv_choise
      TABLES
        valuetab     = lt_path
      EXCEPTIONS
        break_off    = 1
        OTHERS       = 2.
    IF sy-subrc <> 0.
* Implement suitable error handling here
    ENDIF.
  • Step 4: Active all objects and test de functionality. TCode SPRO–> F5 –> “Copy IMG Path”