IP Geolocation Compact — CSV

At a glance

What's in it

IP ranges with LocationId, ConfidenceAreaId, IsCellular, HostingLikelihood, ConfidenceScore and DeploymentExtentKm; locations; confidence areas; place names in 10 languages (German, English, Spanish, French, Italian, Japanese, Korean, Portuguese, Russian and Chinese), compiled and maintained by BigDataCloud. Each language falls back to the English name where a translation isn't published.

What it's for

City-level location for your company's own fraud checks, analytics and reporting.

Other packages add

BGP prefixes, RIR registry ranges and an ASN per IP in Network Data (see Network CSV); hyper-local, in-city location on wired networks, each ASN's most active operational area, and proxy, Tor and VPN lists in Complete (see Complete CSV).

Scale

Tens of millions of IPv4 and IPv6 ranges. Manifest.csv has the exact counts for each release.

Licence

Flat Data Standard Licence: internal use by your company and its employees only.

Files in the archive

Fields

How the files join

Manifest.csv describes the release, and NOTICE.txt holds third-party data attributions.

To geolocate an IP: find the Ranges row where RangeStart ≤ IP ≤ RangeEnd, then join LocationId to Locations.csv (and its GeoNames IDs to the language file you need) and ConfidenceAreaId to ConfidenceAreas.csv. Compact ranges don't carry an ASN; for IP-to-ASN, see Network CSV.

Sample data

Sample rows from the 11 October 2026 release. Values change with each daily update.

Ranges-IPv4.csv: a reserved-space row, a Melbourne range and a cellular range in the Philippines

RangeStart,RangeEnd,LocationId,ConfidenceAreaId,IsCellular,HostingLikelihood,ConfidenceScore,DeploymentExtentKm
0.0.0.0,0.0.1.255,,,False,0,0,
139.130.4.0,139.130.4.31,150626,1742,False,0,3,
1.37.82.0,1.37.82.255,2001,690,True,0,4,250

Ranges-IPv6.csv: a Melbourne range and a range in Rochester, New York

RangeStart,RangeEnd,LocationId,ConfidenceAreaId,IsCellular,HostingLikelihood,ConfidenceScore,DeploymentExtentKm
2001:8000::,2001:8000:fff:ffff:ffff:ffff:ffff:ffff,2319,221465,False,0,4,1000
2001:1890:152a:5b00::,2001:1890:152a:5bff:ffff:ffff:ffff:ffff,59587,221659,False,0,2,

Locations.csv

LocationId,Latitude,Longitude,ContinentIso,CountryIso,SubdivisionIso,SubdivisionGeonamesId,CityGeonamesId,LocalityGeonamesId,Postcode,Timezone
150626,-37.82,145.16,AU,AU,AU-VIC,2145234,2158177,7932624,3130,Australia/Melbourne
2001,14.77,121,AS,PH,PH-03,7521305,1699802,1699802,,Asia/Manila

Joined result for 139.130.4.5 (from the DuckDB example below): range 139.130.4.0–139.130.4.31, AU, AU-VIC, Melbourne, -37.82, 145.16, confidence area 1742. 1.37.82.10 joins to Meycauayan, PH-03, Philippines.

Behaviour and edge cases

Supplied fields

Each row fills the fields that apply to it, and each field description says when it is supplied. Check that a field has a value before you use it.

Private and reserved addresses

Where Ranges-* have rows for private and reserved space (for example 0.0.0.0–0.0.1.255, and the start of Ranges-IPv6.csv), the row carries IsCellular, HostingLikelihood and ConfidenceScore. LocationId and ConfidenceAreaId are supplied where collected network information is statistically significant.

IDs belong to the package

LocationId and ConfidenceAreaId are specific to Compact and to each release, so join them only to lookup files from the same download. GeoNames IDs are the same in every package.

File conventions

UTF-8. The Geonames-*.csv files start with a BOM, so use a BOM-aware reader. Lines end in CRLF in every file, Manifest.csv included. Text fields can be double-quoted. Booleans are True/False. Full list: CSV conventions.

Download

Download from your BigDataCloud account, or automate it with your API key:

curl -fL -o bigdatacloud-compact-csv.zip "https://www.bigdatacloud.com/download?package=compact&format=csv&key=YOUR_API_KEY"

URL=$(curl -s -D - -o /dev/null -H 'x-bdc-key: YOUR_API_KEY' \
  "https://www.bigdatacloud.com/download?package=compact&format=csv" \
  | awk 'tolower($1)=="location:" { sub(/\r$/, "", $2); print $2 }')
curl -fo bigdatacloud-compact-csv.zip "${URL:?no download link: check the key and package}"

Updates and integrity

• Updated daily. Download on any schedule that suits you; updates are included in your subscription.
• Manifest.csv records the build time (Build Complete), the data timestamps and the row totals, then one row per other file in that ZIP, including NOTICE.txt where present, with its byte count, so you can check a complete download and load. To check the ZIP itself, see Check a download.

Code samples

Look up one IP and join to place names (DuckDB, IPv4). DuckDB reads the CSV files directly, so you don't need to load them into a database first.

import duckdb

con = duckdb.connect()
# IPv4 dotted-quad -> integer, so ranges can be compared numerically
con.sql("""
CREATE MACRO ip4(ip) AS
  (split_part(ip,'.',1)::BIGINT << 24) + (split_part(ip,'.',2)::BIGINT << 16)
+ (split_part(ip,'.',3)::BIGINT << 8)  +  split_part(ip,'.',4)::BIGINT
""")
print(con.sql("""
SELECT r.RangeStart, r.RangeEnd, l.CountryIso, l.SubdivisionIso,
       g.Name AS City, l.Latitude, l.Longitude, r.ConfidenceAreaId
FROM read_csv('Ranges-IPv4.csv', all_varchar = true) r
LEFT JOIN read_csv('Locations.csv', all_varchar = true) l USING (LocationId)
LEFT JOIN read_csv('Geonames-en.csv', all_varchar = true) g
       ON g.GeonamesId = l.CityGeonamesId
WHERE ip4('139.130.4.5') BETWEEN ip4(r.RangeStart) AND ip4(r.RangeEnd)
""").fetchall())
# [('139.130.4.0', '139.130.4.31', 'AU', 'AU-VIC', 'Melbourne', '-37.82', '145.16', '1742')]

For IPv6 or high volumes, load the files into your database once, store range bounds as numbers or your database's IP type, and index them.

Related

• Compact MMDB
• Lookup files reference
• Downloads & automation
• Complete CSV · Network CSV
• Overview