Data types used in release files
GeoIP Locations publishes monthly datasets in CSV, Parquet, and MMDB formats.
Format overview
CSV is a text-based tabular format designed for simple exchange and inspection.
Parquet is a columnar analytics format designed for efficient querying and batch processing.
MMDB is an IP lookup database format designed for fast local IP-to-record resolution.
Schema alignment
For normal country, ALL, EDU, and GOV scopes, CSV and Parquet use the same flat release schema. MMDB carries the same release concept and comparable attribution fields, but it is not a flat table. It stores lookup records attached to IP networks.
Important exception: BOG packages use a reduced schema because bogon records do not represent normal geolocation rows.
Standard fields
The standard flat schema includes network, network_from, network_to, prefix_length, country, region, city, latitude, longitude, timezone, ASN, ASN type, network type, company name, RIR, and snapshot month fields.
Recommended usage
Use CSV for simple download and inspection, Parquet for analytics workflows, and MMDB for local lookup services or applications.