French Patent Documents From 2020 To 2026

French patent documents published between 2020 and 2026, with one full-text document per row. The dataset contains over 468,000 records and was created by INPI-France from original XML sources. It was last updated in April 2026.

Use Cases

• Analyze patent text content for trends in specific technology sectors over the 2020-2026 period.
• Train text classification models to categorize patents based on their full-text descriptions.
• Perform named entity recognition on patent documents to extract inventor names, company affiliations, and technical terms.
• Study the evolution of patent language and claims across different publication years within the dataset.

Strengths

• Contains over 468,000 individual patent records.
• Covers a 7-year time range from 2020 to 2026.
• Data is provided in Parquet format suitable for distributed processing.

Limitations

• Specific column names and data structure are not provided, complicating initial analysis.
• The sample size relative to the global patent universe is unknown, potentially limiting generalizability.

Provenance

Source

French Patent and Trademark Office (INPI) public patent documents (A1).

Collection Method

Extracted, structured, and cleaned from original XML sources via API/FTP access.

Time Range

2020 to 2026

Freshness

Data is current through 2026, with the dataset updated in April 2026.

Geography

France