French Patent Documents from 1981 to 2026 in Cleaned Format

French patent documents published between 1981 and 2026, extracted from original XML sources with one document per row. The dataset is provided in Parquet format by INPI-France and was last updated on March 23, —. Each row likely contains the full text of a patent publication.

Use Cases

• Patent trend analysis based on the 45-year publication time range
• Natural language processing on full-text French technical and legal documents
• Intellectual property landscape mapping based on publication data
• Information extraction for inventor, assignee, or technology classification fields likely present in the structured XML

Strengths

• Covers a 45-year time range from 1981 to 2026
• Data is cleaned and structured from original XML sources
• Provided in Parquet format, suitable for distributed or streaming processing

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment
• Description metadata is limited; actual data quality requires manual inspection after download

Provenance

Source

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

Collection Method

Independent extraction, structuring, and cleaning from API/FTP PI access

Time Range

1981 to 2026

Freshness

Last updated 2026-03-23 14:00:06; freshness should be verified

Geography

France