Coming from Microsoft MDS

Grunddata is a cloud-hosted replacement for Microsoft Master Data Services. If you have run MDS you will recognise most of the model, and the Excel workflow will feel familiar.

Concept mapping

Microsoft MDS	Grunddata	Notes
Model	Model	Same idea — a grouping of entities.
Entity	Table	Same idea.
Attribute	Attribute	Same idea, plus JSON Schema validation.
Domain-based attribute	Domain table attribute	Same idea. The referenced table must be MANUAL.
Member	Record	A row.
Code / Name	Code / Name	Built-in on every table, same purpose.
Business rules	Validations	Expressed as JSON Schema, built through a form.
Subscription views	Views	Published as OData rather than SQL views.
Excel add-in	Excel plugin	The main tool for maintaining data.
Transaction log	Commits	Every change is attributed and versioned.
Model versions	(intentionally absent)	See below.

What is deliberately different

Model versioning

MDS lets you version a model. Grunddata does not, and this is on purpose: MDS versioned the data but always applied the latest structure, which produced misleading history.

Instead, if you need to preserve an old structure alongside a new one, copy the model and put the version in its name:

COREP RF 4.0
COREP RF 4.2
COREP RF 4.3

Tip

For most teams this is only needed for externally-mandated frameworks that change on a schedule. Everyday changes are handled by record versioning, which Grunddata does keep in full.

Record history is always on

In Grunddata every record is append-only. There is no configuration to enable history and no purge job to schedule — the version trail is the storage model.

Downstream consumption is OData, not SQL

Instead of subscription views in a database you control, Grunddata publishes views over an OData feed that Excel, Power Query and Power BI consume natively.

Google Sheets is supported

In addition to Excel, Grunddata provides a Google Sheets add-on with the same download / edit / commit workflow.

Migration approach

1. Recreate your MDS models as Grunddata models.
2. Recreate entities as tables. Anything that was used as a domain-based attribute target must be created with Code generation = MANUAL.
3. Recreate attributes, keeping names identical where possible so existing spreadsheets stay familiar.
4. Translate business rules into validations, and use Preview to check them against your data before enforcing them.
5. Load lookups first, then the tables that reference them. Order matters: a domain reference can only resolve if the target record already exists.
6. Create views for anything that had a subscription view, and repoint your reports.