🎯 Definition

A Virtual Catalog Item (VCI) in Ataccama ONE is a logical representation of data stored in the Data Catalog that does not map directly to a single physical table or file. Instead, VCIs utilize data virtualization to combine, filter, or transform metadata and columns from one or more physical sources (like databases or APIs) into a single, logical virtual table, without physically replicating or moving the underlying data (zero-copy data sharing).

🔑 Key Takeaways

  • Zero-Copy Access: VCIs expose structured views of data to users without running expensive ETL replication or consuming additional physical storage.
  • Semantic Layering: VCIs act as a bridge between technical databases (with complex schemas and cryptic column names) and business-friendly representations.
  • Virtual Joins: Allow joining data across different sources (e.g., joining an Oracle customer table with a CSV file on a secure share) dynamically at query time.

📚 Detailed Explanation

In traditional catalog architectures, every catalog item is a direct reflection of a physical database asset (such as an oracle_db.customers table).

Ataccama ONE introduces VCIs to support agile data consumption and governance:

  1. Logical Views: Similar to a view in SQL, a VCI is defined by a query or a set of transformation rules (built visually or via SQL expressions).
  2. Governance Layer: Because VCIs are native Catalog items, they inherit standard Ataccama ONE governance features:
    • They can have automated Data Quality Rules assigned to them.
    • They can be mapped to business terms in the Glossary.
    • Access permissions can be controlled at the VCI level (restricting access to specific virtual columns).
  3. Performance: At query time, Ataccama’s virtualization engine translates the request into optimal queries pushed down (Query Pushdown) to the source databases, aggregating results on-the-fly.

💡 Use Cases & Examples

  • Data Masking for Analytics: A data steward creates a VCI based on a physical Employees table but excludes sensitive columns (e.g., Social_Security_Number, Salary). The data science team is granted access to the VCI instead of the raw physical table.
  • Federated Customer Profile: A VCI that virtualizes and joins transactional data from PostgreSQL and reference country codes from a CSV file to present a unified “Customers with Countries” object in the catalog search interface.

📌 Sources