Glossary/Data Governance & Quality

Technical Metadata

Technical metadata is information about the structural and technical properties of data, including schema, data types, lineage, storage location, and refresh schedules.

Technical metadata describes how data is structured and managed. It includes: table and column names, data types, constraints, cardinality, lineage (which systems feed which tables), refresh schedules, storage location, compression method, and performance characteristics. Technical metadata is often auto-discoverable from source systems: data warehouses publish schemas, data pipelines log lineage, monitoring systems capture refresh times.

Technical metadata is essential for data engineers and analysts who need to understand technical implementation. It answers questions: Is this column an integer or string? How frequently is this table updated? Which upstream system feeds this table? Can I join these two tables on customer ID? Without technical metadata, engineers spend time investigating systems; with it, answers are available in a catalog.

Technical metadata is typically maintained automatically through schema inference, lineage tracking, and system integration. Tools like dbt infer lineage from SQL; cloud data warehouses auto-discover schemas; data quality tools capture freshness and error information. The value of technical metadata increases when it's automated because it stays current as systems change.

Key Characteristics

  • Describes data structure and organization
  • Includes schemas, data types, and constraints
  • Captures lineage and dependencies
  • Tracks refresh schedules and latency
  • Often auto-generated from systems
  • Version-controlled with historical tracking

Why It Matters

  • Navigation: Engineers understand data structure without investigation
  • Quality: Metadata about refresh schedules and lineage enable validation
  • Performance: Schema and cardinality information guide optimization
  • Reliability: Understanding dependencies enables impact analysis
  • Automation: Technical metadata enables automated data quality checks

Example

Technical metadata for an orders table: type = fact table, columns = order_id (int, PK), customer_id (int, FK), amount (decimal), created_at (timestamp), source = production database, refresh = hourly, lineage = [transaction_source -> staging_orders -> orders], size = 500GB.

Coginiti Perspective

Coginiti works with technical metadata across its 24+ platform connectors, accessing each database's schema, type, and constraint information. CoginitiScript blocks can query system catalogs and information_schema tables to extract technical metadata programmatically. SMDL entities that map to tables inherit the technical metadata of their source, while query-based entities document their derivation logic. The object store browser surfaces file-level technical metadata (format, size, location) for data on S3, Azure Blob, and GCS.

See Semantic Intelligence in Action

Coginiti operationalizes business meaning across your entire data estate.