Documentation

How to use this data

Use ProdIPData releases as local, versioned GeoIP intelligence. The core workflow is straightforward: choose a release month, load CSV or Parquet files, preserve the snapshot month, and query by numeric IPv4 prefix.

These examples assume the public release layout with monthly folders, ISO-scoped packages, and a numeric prefix column representing the first IPv4 address in a /24 block.

Open downloads Accuracy and limitations

Operating model

Choose the right format first

CSV is the most universal integration format. Parquet is the better default for analytical processing, notebooks, DuckDB, and columnar warehouse-style workflows.

CSV

Use CSV for SQL Server bulk imports, operational review, simple ETL flows, and teams that prefer transparent row-oriented files.

Parquet

Use Parquet for DuckDB, analytics pipelines, notebooks, fast column scans, and large release-level aggregations.

MMDB

Use MMDB when the application needs embedded lookup behavior rather than tabular analytics or warehouse-style joins.

Core lookup concept

Convert the dotted IPv4 address to a numeric integer, then locate the row where the address falls inside the published prefix range.

-- Conceptual lookup model
ip_integer >= prefix
AND ip_integer < prefix + 256

This model works because a /24 row covers 256 addresses. For example, the row starting at 134744064 covers the range 8.8.8.0 through 8.8.8.255.

Recommended table practices

Keep snapshot_month in every table, index, export, and downstream view.
Use BIGINT for numeric IPv4 values to keep arithmetic clear and portable.
Index prefix, iso, asn, and snapshot_month according to query pattern.
Load one release into a staging table before promoting it to a production lookup table.

SQL Server

Load and query CSV release files

SQL Server is a strong fit when the dataset becomes part of existing reporting, reconciliation, or production-enrichment workflows.

1. Create a landing table

CREATE TABLE dbo.ProdIP_Prefix_Attributes
(
    prefix          BIGINT        NOT NULL,
    city            NVARCHAR(200) NULL,
    region_state    NVARCHAR(200) NULL,
    country         NVARCHAR(200) NULL,
    iso             CHAR(2)       NULL,
    latitude        DECIMAL(9,6)  NULL,
    longitude       DECIMAL(9,6)  NULL,
    timezone        NVARCHAR(128) NULL,
    asn             NVARCHAR(300) NULL,
    company         NVARCHAR(300) NULL,
    host            NVARCHAR(300) NULL,
    snapshot_month  CHAR(7)       NOT NULL
);

CREATE INDEX IX_ProdIP_Prefix_Attributes_Prefix
ON dbo.ProdIP_Prefix_Attributes(prefix);

CREATE INDEX IX_ProdIP_Prefix_Attributes_Snapshot_Iso
ON dbo.ProdIP_Prefix_Attributes(snapshot_month, iso);

2. Bulk load a decompressed CSV file

If the downloaded CSV is compressed, unzip and decompress it first, then point BULK INSERT to the plain CSV path.

BULK INSERT dbo.ProdIP_Prefix_Attributes
FROM 'D:\SCRAP\Export\CSV\release_2026-03\iso=US\prefix_attributes_2026-03_iso-US_part-0001-of-0001.csv'
WITH
(
    FORMAT = 'CSV',
    FIRSTROW = 2,
    FIELDQUOTE = '"',
    CODEPAGE = '65001',
    TABLOCK
);

3. Convert dotted IPv4 to integer

CREATE OR ALTER FUNCTION dbo.IPv4ToBigInt
(
    @ip VARCHAR(15)
)
RETURNS BIGINT
AS
BEGIN
    DECLARE
        @a BIGINT = TRY_CONVERT(BIGINT, PARSENAME(@ip, 4)),
        @b BIGINT = TRY_CONVERT(BIGINT, PARSENAME(@ip, 3)),
        @c BIGINT = TRY_CONVERT(BIGINT, PARSENAME(@ip, 2)),
        @d BIGINT = TRY_CONVERT(BIGINT, PARSENAME(@ip, 1));

    IF @a IS NULL OR @b IS NULL OR @c IS NULL OR @d IS NULL
       OR @a NOT BETWEEN 0 AND 255
       OR @b NOT BETWEEN 0 AND 255
       OR @c NOT BETWEEN 0 AND 255
       OR @d NOT BETWEEN 0 AND 255
    BEGIN
        RETURN NULL;
    END;

    RETURN (@a * 16777216) + (@b * 65536) + (@c * 256) + @d;
END;
GO

4. Lookup one IP address

DECLARE @ip BIGINT = dbo.IPv4ToBigInt('8.8.8.8');

SELECT TOP (1)
    prefix,
    city,
    region_state,
    country,
    iso,
    latitude,
    longitude,
    timezone,
    asn,
    company,
    host,
    snapshot_month
FROM dbo.ProdIP_Prefix_Attributes
WHERE snapshot_month = '2026-03'
  AND @ip >= prefix
  AND @ip < prefix + 256
ORDER BY prefix DESC;

5. Aggregate country coverage

SELECT
    iso,
    country,
    COUNT_BIG(*) AS prefix_count,
    COUNT_BIG(*) * 256 AS ipv4_count,
    COUNT(DISTINCT asn) AS observed_asn_count
FROM dbo.ProdIP_Prefix_Attributes
WHERE snapshot_month = '2026-03'
GROUP BY iso, country
ORDER BY ipv4_count DESC;

DuckDB

Analyze Parquet release files locally

DuckDB is excellent for local analysis because it can query Parquet directly without a separate database server or import cycle.

1. Create a view over Parquet files

CREATE OR REPLACE VIEW prodip_release AS
SELECT *
FROM read_parquet('D:/SCRAP/Export/PARQUET/release_2026-03/iso=*/prefix_attributes_2026-03_iso-*.parquet');

2. Lookup one IP address

WITH ip_input AS (
    SELECT
        CAST(split_part(ip, '.', 1) AS UBIGINT) * 16777216 +
        CAST(split_part(ip, '.', 2) AS UBIGINT) * 65536 +
        CAST(split_part(ip, '.', 3) AS UBIGINT) * 256 +
        CAST(split_part(ip, '.', 4) AS UBIGINT) AS ip_int
    FROM (VALUES ('8.8.8.8')) AS v(ip)
)
SELECT
    r.prefix,
    r.city,
    r.region_state,
    r.country,
    r.iso,
    r.latitude,
    r.longitude,
    r.timezone,
    r.asn,
    r.company,
    r.host,
    r.snapshot_month
FROM prodip_release r
CROSS JOIN ip_input q
WHERE q.ip_int >= r.prefix
  AND q.ip_int < r.prefix + 256
LIMIT 1;

3. Build a country summary

SELECT
    iso,
    country,
    COUNT(*) AS prefix_count,
    COUNT(*) * 256 AS ipv4_count,
    COUNT(DISTINCT asn) AS observed_asn_count
FROM prodip_release
GROUP BY iso, country
ORDER BY ipv4_count DESC;

4. Identify large ASN footprints

SELECT
    asn,
    COUNT(*) AS prefix_count,
    COUNT(*) * 256 AS ipv4_count,
    COUNT(DISTINCT iso) AS country_count
FROM prodip_release
WHERE asn IS NOT NULL
  AND asn <> 'n/a'
GROUP BY asn
ORDER BY prefix_count DESC
LIMIT 50;

5. Query CSV directly when needed

SELECT *
FROM read_csv_auto(
    'D:/SCRAP/Export/CSV/release_2026-03/iso=US/*.csv.gz',
    header = true
)
LIMIT 10;

Common join pattern

The cleanest integration pattern is to enrich your internal event table with a computed numeric IPv4 column, then join to the release table by range.

SELECT
    e.event_id,
    e.ip_address,
    g.country,
    g.iso,
    g.city,
    g.asn,
    g.company,
    g.snapshot_month
FROM dbo.Events e
JOIN dbo.ProdIP_Prefix_Attributes g
  ON e.ip_integer >= g.prefix
 AND e.ip_integer < g.prefix + 256
WHERE g.snapshot_month = '2026-03';

Production rollout checklist

Download a specific release month and archive the original files.
Load into staging and validate row counts before promotion.
Preserve release month in table names, partitions, or metadata.
Run sample lookups for known infrastructure IPs.
Publish an internal data contract that explains the accuracy limits.