Common Access Pattern - GraphQL and Filedownload

This page defines one reusable access strategy for Grunddatamodellen datasets.

Use this as a shared pattern for CVR, DAR, Matrikel, DAGI, and related registries.

Decision Table

Access mode	Best when	Typical output shape	Cost profile
GraphQL	Hard filters known early (status, municipality, code lists, time windows)	Targeted record subsets	Lower post-processing, higher query design effort
Filedownload	Broad extraction/snapshot needed	One file per normalized entity	Fast bulk extraction, higher join and reconciliation effort

Shared Workflow

Step 1: Choose temporal mode first

Before query design, declare one mode:

• current
• temporal
• bitemporal

See Shared Temporal Superclass Contract.

Step 2: Select extraction mode

• GraphQL for selective extraction
• Filedownload for broad extraction

Step 3: Normalize to analysis model

Grunddata entities are heavily normalized.

A general reconstruction pattern is:

1. choose anchor entity
2. attach relation entities by shared IDs
3. resolve one-to-many expansions
4. publish analysis-ready view(s)

Step 4: Run quality gates

• temporal-mode declaration is present
• join coverage/match rate is measured
• duplicates introduced by relation joins are reported
• status filter assumptions are documented

AI-Facing Contract

Machine-readable version:

• grunddata.common.access-pattern.v1.json
• datafordeler.entity-filedownload.pattern.manifest.v1.json

Dataset-specific playbooks should inherit this pattern and only add domain-specific overrides.