EBA欧洲银行-DPM-Database-description_20页_476kb

Data Point Model Database Summary

Introduction

The Data Point Model (DPM) database was introduced by the EBA in August 2013 to enhance the implementation of the Data Point Modelling methodology. It replaces the earlier MS Excel-based approach with a relational database using MS Access, allowing for logical constraints, automatic consistency checks, and improved metadata management. This helps streamline the process of achieving high-quality data models, especially for DPMs that categorize over 30,000 data points. The database is a meta-model, focusing on core reporting concepts such as framework, table, dimension, member, and domain, and is designed to be adaptable to various reporting domains, including but not limited to COREP and FINREP.

Core Concepts and Structure

The DPM database is structured around:

• Reporting Frameworks (e.g., COREP, FINREP)
• Tables (which may be grouped into TableGroups and subject areas)
• TableVersions (representing changes in the structure of a table over time)
• Modules (grouping related tables into a single submission unit)

Tables and Table Groupings

• Tables are part of a reporting framework and can be closed (fixed size) or open (variable size).
• Closed tables have all values explicitly listed, while open tables may have open axes (X, Y, or Z) which allow for variable entries.
• TableGroups are used to organize related tables, especially those that form parts of a larger business template.
• Open tables can be of two types: lists (open Y-axis) and tables with optional sheets (open Z-axis).

Table Layout and Axes

• Axes represent rows, columns, or sheets in a table. Each axis has an AxisOrientation (X, Y, Z).
• AxisOrdinates are specific positions on an axis. They are used to define the structure of a table and can be open or closed.
• Open axes are marked by the IsOpenAxis field and use a special code "999" to represent unknown values.
• Tree structures are represented in the model using AxisOrdinates, HierarchyNodes, and TableGroups. Each node has a Parent and a Path to indicate its position in the hierarchy.

Data Point and Dimensional Categorization

• A DataPoint represents a unique item of business information, independent of its categorization in a specific taxonomy.
• DataPointVersions track the categorization of a DataPoint over time, including its metric, context, and valid dates.
• Dimensions are categories used to describe data points, with members being the specific values within those dimensions.
• Domains define the allowable values for a dimension. Some domains are closed (e.g., Countries), while others are open (e.g., Legal Entities).
• Hierarchies define relationships between members and support aggregation from lower to higher levels.

Dimensional Coordinate and Categorization

• A DimensionalCoordinate is a pair of [Dimension].[Member] that uniquely identifies a data point.
• The categorization of a data point is derived from the combination of dimensional coordinates from its axis ordinates.
• Consistency checks are performed iteratively to ensure logical integrity, identifying issues such as missing mandatory dimensions or duplicate dimensions.

Validation and XBRL Integration

• Validation rules are embedded in the DPM database and can be referenced in the accompanying "Eurofiling - Validations.pdf".
• The DPM is designed to be XBRL-ready, with Metric mapping directly to primary items in XBRL.
• ContextDefinitions and ContextOfDataPoints are used to map the dimensional categorization to XBRL contexts.
• Future enhancements will include XBRL properties such as namespaces and versioning of metadata to support time series analysis and data warehousing.

Key Features and Benefits

• Logical constraints and automatic consistency checks improve model quality and reduce manual effort.
• Multiple views of metadata support diverse user needs.
• Open tables are supported with special handling for variable entries and repeating units.
• Tree structures are represented using adjacency lists and path enumeration for flexibility and clarity.
• DataPointVersions allow tracking of metadata changes over time.

Conclusion

The Data Point Model database is a comprehensive and flexible tool for managing and validating reporting metadata. It supports both closed and open tables, dimensional analysis, and XBRL integration. The model is designed to be iterative, adaptable, and user-focused, enabling efficient and accurate data reporting across different frameworks and systems.