Metadata Repository

This epic enables automated analysis of relational source databases and stores the results in a centralized metadata repository. It helps customers understand database structures, detect sensitive data, identify relationships, and use this foundation for safer ETL processing and more reliable synthetic test data generation. The repository provides a consistent basis for traceability, rule-based processing, and reusable data intelligence across ETL use cases.

Details
Epic
XEP-5 - Metadata Repository
Status
Approved
Progress

12 Stories - 0 completed - 12 in progress

0% completed

Problem Statement

When source databases are unknown or poorly documented, ETL teams must manually inspect structures, relationships, and sensitive fields before they can process data reliably.

This slows down onboarding, increases the risk of errors, and makes it difficult to extract related data consistently, protect confidential information, and define realistic rules for synthetic data generation.

Customer Value

Customers can onboard unfamiliar databases faster and reduce manual analysis effort. They gain a reusable metadata repository that supports consistent ETL decisions across discovery, masking, extraction, and documentation.

This same repository also provides the foundation for synthetic test data generation by capturing the structural, semantic, and statistical rules needed to create realistic but non-personal data.

As a result, customers can improve data protection, maintain consistency across related datasets, and accelerate testing and development.

Scope

This epic covers the creation and use of a centralized metadata repository derived from relational databases, including:

  • Automatic discovery of tables, columns, data types, constraints, and indexes

  • Storage of technical metadata in a structured, reusable repository

  • Profiling of tables and columns, such as null rates, value ranges, distinct counts, frequent values, and detected patterns

  • Automatic classification of columns based on semantic meaning, including sensitive or business-relevant fields

  • Detection and assessment of existing or likely relationships between tables

  • Marking sensitive fields to support masking decisions

  • Using detected relationships to enable consistent extraction of related data across multiple tables

  • Using stored profiles, classifications, and relationships to derive rules for synthetic test data generation

  • Providing evidence, confidence, and status for automated findings

  • Supporting review workflows so results can be confirmed, corrected, or rejected manually

  • Improving documentation and traceability of database structures and automated decisions

Stories
Story Title Release Status
XPU-3137 MR-01 – Save Metadata Catalog and Display It Under a Connection Planned
XPU-3138 MR-02 – Read Technical Metadata via JDBC Planned
XPU-3139 MR-03 – Configure the “Metadata Scan” Task Type Planned
XPU-3140 MR-05 – Implement a Batch Task Template for Metadata Scans Planned
XPU-3141 MR-06 – Capture Table Structures, Constraints, and Indexes Planned
XPU-3142 MR-08 – Run Configurable Data Profiling Planned
XPU-3143 MR-09 – Detect Data Formats and Patterns Planned
XPU-3144 MR-10 – Suggest Semantic Classes and Sensitivity Planned
XPU-3145 MR-11 – Detect Implicit Table Relationships Planned
XPU-3146 MR-12 – Display Analysis Results in the Metadata Catalog Planned
XPU-3147 MR-13 – Review Classifications and Relationships Planned
XPU-3148 MR-14 – Track Scan History and Structural Changes Planned

UBS Hainer GmbH - generated from Jira at Tue, 3 Nov 2026 13:29:07 GMT