MDG Materials Duplicate Check

[stefan_bogesch]

Hello experts,

we have configured duplicate check in MDG-M on Material description (medium). It works fine, but sometimes there is a strange behavouir on the duplicate check.

The duplicate check pops up, even though there is no duplicate.

Does anybody knows this issue or knows how to resolve?

[MOHAMMED_BASER · Accepted solution]

Dear @stefan_bogesch 
The reason why the Duplicate Check Pops Up Without Real Duplicates due to following reasons. Fuzzy Matching Logic, SAP MDG uses Data Quality Management (DQM) or generic duplicate check methods that rely on similarity scores. If two descriptions are “close enough” (e.g., Pump Motor vs. Pump-Motor), the system flags them. Normalization Rules, the duplicate check often strips out spaces, hyphens, or capitalization. For example, Steel Rod and STEELROD may be treated as duplicates.

Indexing Delays or Cache Effects also. Sometimes the search index isn’t fully updated, so older entries or partial matches cause false positives.

Threshold Settings for the duplicate check has configurable thresholds for similarity. If the threshold is set too low, even minor overlaps trigger the popup.

Search Provider Behavior, Depending on whether you use Enterprise Search or HANA Search, the matching logic differs. Enterprise Search tends to be broader, while HANA Search can be tuned more precisely.

The best practice is to combine Material Description with other attributes (like Material Type or Group) for duplicate checks. Regularly cleanse material master data to avoid near-duplicates that confuse the algorithm. Document your match profiles and thresholds so the team understands why certain cases trigger warnings.

[stefan_bogesch · Accepted solution]

Dear @MOHAMMED_BASER,

thanks for your response, that helped me to understand it better.

We use HANA search and have the thresholds set to 50 and 80.

So would it help to set them higher? Do both need to be set higher?

Regards

Stefan

[MOHAMMED_BASER · Accepted solution]

Hi, I’m glad the explanation helped clarify things. Regarding your question on thresholds: In HANA search, the lower threshold (50) typically defines when a potential duplicate candidate is considered for comparison, while the upper threshold (80) defines when the system flags it as a strong match. If you set these thresholds higher, the system becomes stricter, meaning fewer records will be flagged as duplicates. This can reduce false positives, but it also increases the risk of missing genuine duplicates. Best practice is to adjust them gradually and test with real data. For example, raising the lower threshold from 50 to 60 may already reduce unnecessary popups without losing too many true matches. You don’t necessarily need to raise both thresholds equally. Often, tuning the lower threshold has the biggest impact on reducing false positives, while the upper threshold ensures only strong matches trigger warnings. I’d recommend experimenting in a test environment with different values (e.g., 60/85) and reviewing the results with your data team before applying changes in production. Regards, Mohammed Baser