Reference22 Reference · 2 of 3
Glossary
Platform vocabulary, plus the German terms a DACH client will use in the room. If you are coaching in English on a German project, the right-hand column is the one that prevents an hour of talking past each other.
Stand 29.07.2026
| Term | Means | German |
|---|---|---|
| Bronze / Silver / Gold | Refinement stages. Not a data model. | Rohschicht / Bereinigungsschicht / Auswertungsschicht |
| Grain | What exactly one row of a table represents. | Granularität |
| Fact table | Measurements at a declared grain. | Faktentabelle |
| Dimension | The descriptive context you group and filter by. | Dimension |
| Surrogate key | A system-generated key standing in for the business key. | Surrogatschlüssel |
| Business key | The identifier the source system uses. | Geschäftsschlüssel |
| SCD1 | Overwrite on change; no history kept. | Historisierung Typ 1 |
| SCD2 | A new row per change, with validity intervals. | Historisierung Typ 2 |
| Conformed dimension | One dimension shared by several facts. | Konforme Dimension |
| Unknown member | The placeholder row a fact points at when the dimension is missing. | Unbekannt-Eintrag / Dummy-Satz (häufig unübersetzt: Unknown Member) |
| Data Vault | Hub / Link / Satellite modelling. Common in German enterprises. | Data Vault |
| Hub / Link / Satellite | Business key / relationship / descriptive history. | Hub / Link / Satellit |
| Idempotent | Running it twice leaves the same result as running it once. | Idempotent |
| CDC | Change data capture — a stream of source changes. | Änderungsdatenerfassung (CDC) |
| Full extract | The whole source, re-delivered each time. | Vollabzug |
| Delta / incremental load | Only what changed since last time. | Delta-Beladung / inkrementelle Beladung |
| Snapshot comparison | Diffing today against yesterday when the source has no change timestamp. | Snapshot-Vergleich |
| Auto Loader | Incremental file ingestion with schema evolution. Two words, capitalised — that is how Databricks writes it. | Auto Loader |
| Expectation | A declared data-quality rule on a pipeline: warn, drop or fail. | Erwartung / Qualitätsregel |
| Lakeflow pipeline | Declarative pipeline; formerly Delta Live Tables. | Lakeflow-Pipeline |
| Streaming table | A table maintained incrementally from a stream: appends, and — via AUTO CDC — upserts and deletes as well. 'Append-only' is the old, narrower reading. | Streaming-Tabelle |
| Materialized view | A precomputed, refreshable query result. Not a metric view. | Materialisierte Sicht |
| Metric view | A governed semantic definition of measures and dimensions. | Kennzahlensicht |
| MEASURE() | The SQL function that reads a measure from a metric view. | — |
| Unity Catalog | The governance control plane: metastore, catalogs, grants, lineage. | Unity Catalog |
| Metastore | The top-level UC container, one per region. | Metastore |
| Volume | A UC-governed location for non-tabular files — the landing zone Auto Loader reads. Nothing to do with data volume. | Volume (unübersetzt) — nicht „Datenvolumen“ |
| Governed tag | An account-level key-value applied to securables, used in policies. | Governed Tag (unübersetzt) — nie „verwalteter Tag“: Tag heißt hier Schlagwort, nicht Kalendertag |
| ABAC | Access decided by attributes (tags) rather than per-table rules. | Attributbasierte Zugriffskontrolle |
| Row filter / column mask | Policies restricting which rows and columns a principal sees. | Zeilenfilter / Spaltenmaske |
| Liquid clustering | Adaptive data layout replacing partitioning and Z-order. | Liquid Clustering |
| Predictive optimization | Automatic OPTIMIZE, VACUUM and ANALYZE on managed tables. | Predictive Optimization (unübersetzt, so auch im deutschen Databricks-Material) |
| Managed table | A table whose lifecycle and storage UC owns. | Verwaltete Tabelle |
| Deletion vector | A Delta marker that hides deleted rows instead of rewriting the file. The bytes stay readable via time travel until REORG … APPLY (PURGE) and a VACUUM past the retention window. | Löschvektor |
| Data Quality Monitoring | Anomaly detection at catalog/schema level plus data profiling at table level; formerly Lakehouse Monitoring. Anomaly detection is Public Preview and writes to system.data_quality_monitoring.table_results. | Datenqualitätsüberwachung / Anomalieerkennung |
| Delta Sharing | Governed sharing without copying data out. | Delta Sharing |
| Lineage | The recorded derivation path between tables and columns. | Datenherkunft |
| Pseudonymisation | Reversible de-identification. Still personal data under GDPR. | Pseudonymisierung |
| Anonymisation | Irreversible; out of GDPR scope. Hashing is NOT this. | Anonymisierung |
| Personal data | Data relating to an identifiable person. | Personenbezogene Daten |
| Definition of done | The checkable list a change satisfies before merge. | Definition of Done |
| Runbook | One page telling whoever was woken what to do. | Runbook / Betriebshandbuch |
| Characterization test | A test pinning current behaviour before a refactor. | Charakterisierungstest |