xBRL-CSV packages follow a specific structure defined by the OIM specification and EBA filing rules. This guide covers every component of a valid package to help you avoid common structural errors.
Key Takeaways
- Packages are ZIP files with defined structure
- JSON metadata describes the report
- CSV files contain the actual data
- File naming must follow conventions
- Validation checks structure first
Package Overview
A valid xBRL-CSV package:
report.zip
├── report.json # Report metadata
├── facts/
│ ├── table1.csv # Data tables
│ └── table2.csv
└── parameters.json # Optional parameters
JSON Metadata
The report.json file contains:
• Document information
• Entity identifier
• Reporting period
• Taxonomy references
• Table declarations
This metadata enables validation without parsing all CSV files.
CSV Data Files
CSV files contain your reported facts:
• One file per reporting table
• Header row with dimension names
• Data rows for each fact
• UTF-8 encoding required
• Standard CSV delimiters
File Naming
EBA requires specific file naming conventions. Non-compliant names cause immediate rejection. Follow the pattern specified in the EBA filing rules documentation. Case sensitivity matters.