Introducing LakeFusion 5.0

Master data is only as good as the weakest link behind it.

An ungoverned code list, a relationship that never got modeled, a schema that's too rigid for real data. Any one of these can quietly break your golden record. LakeFusion 5.0 tackles all three. It's our biggest release yet, and it's built to give data stewards and architects more governance, more context, and more flexibility across the entire master data management lifecycle, natively on Databricks.

‍

Here's what's new:

1. Reference Data Management

Reference data (think specialty codes across healthcare systems) is the foundation every master record sits on top of, but it rarely gets governed with the same rigor as the master data itself. 5.0 introduces a dedicated Reference Entity type that acts as a true single source of truth for these code lists. You get configurable governance (source wins, steward wins, or flag for review), automated or reviewed record handling, and exact or fuzzy matching. Unresolved codes go straight to a steward review screen so reference data quality stops being a bottleneck for everything downstream.

‍

2. Relationship Management

Entity resolution creates clean golden records; relationship management connects them into a true view of your business. In 5.0, you can model relationships between entities independently, including unidirectional, bidirectional, and any-cardinality relationships (1:1, 1:N, N:N). You can also add attributes to the relationship itself, such as ownership percentage. A new Graph view displays this network natively, while an AI assistant identifies whether you're asking a lookup or analytical question and returns charts, diagrams, or graph expansions directly in chat. This is the difference between having master data and having a real 360-degree view of it.

‍

3. Complex Data Structures

Master data domains almost never fit neatly into flat rows and columns. With native support for structs and arrays, you can model nested objects (a multi-line address) and multi-value attributes (a list of service lines or languages) without flattening anything or duplicating records. Array survivorship handles deduplication automatically and hands back a clean array of distinct values, not a fragile concatenated string, so what comes out the other end is a structurally sound golden record.

‍

4. Deterministic Match Enhancements

Match rule order has always mattered for deterministic match and merge, and now it's a lot easier to get right. Drag and drop rule reordering lets stewards adjust match logic directly instead of rebuilding configuration from scratch, so entity resolution can keep up as source systems and data quality change.

‍

5. Attribute Display Order + NOT Operator

Two smaller changes that came straight from practitioners: configurable attribute display order for cleaner stewardship screens, and a NOT operator for more precise match and governance rules.

‍

Together, these updates advance LakeFusion across the MDM stack: reference data governance, relationship and hierarchy modeling, complex schema support, and more precise matching. The result is golden records you can trust.

‍

Ready to see LakeFusion 5.0 for yourself? Request a demo.

‍