Indu Learning Platform
Job Prep/Example posting
Example prep · 2026-11-01

Senior SAP MDG Technical Consultant

Hub MDG on S/4HANA for a global manufacturer. Builds BP and Material governance with BRFplus, BAdIs and DRF, plus AI-assisted duplicate checks on SAP BTP.

You already show
SAP MDG Business Partner & MaterialData migration (Migration Cockpit, LSMW)DRF replication & hypercareWRICEF specifications
Build before the interview
ABAP OO / BAdI developmentRAP & OData servicesPython for data profilingLLM integration on BTP

Questions

01A change request approved by finance never reached the receiving ECC system. How do you investigate?Scenario·Medium+

First confirm the CR is really activated: final status, data in the active area. Then check whether DRF selected it: run DRFOUT for that BP with the replication model and read DRFLOG. If DRF created a message, follow it to WE02 (IDoc) or SRT_MONI (SOAP) and read the status. Common causes are a DRF filter excluding the business system, missing key mapping, or a value mapping error rejected by the receiver.

  • Activation first
  • DRF filters and replication model
  • IDoc / SOAP monitor
  • Key and value mapping
Step through · 1 of 5
CR activated?DRF selected it?Message created?Receiver statusFix mapping / filter
Check the CR status and that the record is in the active area.
Work down the replication path until it stops.
02Where would you place a validation that blocks a supplier when it has open purchase orders?Design·Medium+

BRFplus can't easily read transactional purchasing data, so this belongs in a BAdI validation. Read open POs through a released CDS view, raise an error message with the PO count, and keep the logic fast enough for mass changes.

  • BRFplus for field rules
  • BAdI for cross-object reads
  • Released CDS views
  • Clear message to requester
BRFplusField checksBusiness-readableBAdIReads PO dataCross-object logicvs
This rule needs data BRFplus can't reach.
03How would you use AI to speed up duplicate detection for business partners without risking wrong merges?AI·Hard+

Keep deterministic rules for tax IDs, add fuzzy and embedding similarity for names and addresses inside blocks, and have an LLM write a short explanation per candidate pair. Stewards approve merges through MDG consolidation or change requests, and acceptance rate is tracked to tune thresholds.

  • Blocking + scoring
  • Embeddings for names
  • LLM explanation
  • Steward approval
  • Acceptance rate
PairScoreAI reasonStewardAcme Pvt Ltd / ACME Private Limited0.96Same tax ID and PINMergeAcme Pvt Ltd / Acme Tools0.71Different tax IDKeep both
AI suggests, the steward decides.

Not in the course yet

BTP CAP vs RAP for side-by-side MDG extensions

The posting mentions side-by-side AI services on BTP.

Lesson generation is turned off on the example. Run your own prep to use it.