datapackage

What it is and what it does

Datapackage-py is a library for working with Data Packages, a specification from Frictionless Data that standardizes how tabular and non-tabular datasets are packaged with metadata. It provides a Package class for managing data package descriptors, a Resource class for reading and iterating over individual datasets, and utilities for validating descriptors against profiles and inferring metadata from raw data files. The library integrates with Table Schema and handles CSV, JSON, and other tabular formats through its runtime dependencies including tableschema and dataflows-tabulator.

The package is designed for data practitioners who need to create reproducible, machine-readable dataset descriptions—common in open data initiatives, data publishing workflows, and data integration pipelines. It lets you load existing data packages from local or remote sources, inspect and modify their descriptors, validate them against the Frictionless specification, and save them as portable archives. However, the project is in dormant maintenance (last release 885 days ago), and the maintainers have released a successor framework; new projects should evaluate whether the Frictionless Framework is a better fit.

Use it for:

• Create standardized metadata for CSV or tabular datasets to publish as open data with machine-readable schemas.
• Infer and validate data package descriptors from raw files to enforce data quality and consistency.
• Load and read tabular resources from a data package descriptor with automatic type casting and missing-value handling.
• Integrate tabular datasets into data pipelines by using Package and Resource classes to handle descriptor-driven data loading.
• Validate foreign key relationships and data integrity constraints defined in a data package descriptor.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides classes and utilities to create, load, validate, and manipulate Data Packages—standardized containers for tabular and non-tabular datasets with machine-readable metadata.

Yes, if you are working within the Frictionless Data ecosystem or need to read/write standardized Data Package descriptors. The low install friction and permissive license make it accessible. However, dormant maintenance (885 days since last release) and outdated Python version classifiers (2.7–3.7) are concerns for new projects; evaluate the Frictionless Framework as a potential alternative if you need active support and modern Python compatibility.

Install

datapackage on PyPI

pip install datapackage

uv add datapackage

poetry add datapackage

Installing datapackage

Before you install

Low install friction with a pure-Python wheel distribution. Maintenance is dormant—last release was 885 days ago—though the repository remains active. The package is in Beta status and supports Python 2.7 through 3.7 per classifiers, which may limit compatibility with modern Python versions.

License in practice

MIT license permits commercial and private use with minimal restrictions, requiring only attribution and inclusion of the license text.

Quickstart

pip install datapackage

from datapackage import Package

package = Package('datapackage.json')
resource = package.get_resource('resource')
data = resource.read(keyed=True)

The package's Python version support (2.7–3.7 per classifiers) may not align with modern Python environments; verify compatibility with your target Python version.

Verify before relying

• Whether the package works reliably with Python versions beyond 3.7 despite classifier declarations.
• Current maintenance status and whether the Frictionless Framework migration path is recommended for new projects.
• Performance characteristics when handling large datasets or complex schemas.

Package facts

Evidence: datapackage-1.15.4-py2.py3-none-any.whl

Keywords: frictionless data, open data, json schema, table schema, data package, tabular data package

Tags

More Utilities packages