UN/LOCODE

The United Nations Code for Trade and Transport Locations: five-character codes for ports, airports, rail and road terminals, cities and other places used in international trade and transport. The code list is maintained by UNECE.

Column definitions and types are in datapackage.json.

Data

• data/code-list.csv: one row per location entry.
• data/subdivision-codes.csv: ISO 3166-2 subdivision codes and names used by the Subdivision column.
• data/country-codes.csv: country codes and names as used in the code list.
• data/alias.csv: alternative names of countries.
• data/function-classifiers.csv: the meaning of each position of the Function column.
• data/status-indicators.csv: the meaning of the Status column.

Columns of data/code-list.csv:

Column	Meaning
Change	Change indicator: + newly added, | changed, # name changed, ! US location with a duplicate IATA code, X to be removed. Empty for entries that did not change
Country	ISO 3166 alpha-2 country code, for example AE. Defined in country-codes.csv
Location	3-character location code (letters and the digits 2 to 9), for example DXB. With Country it forms the 5-character UN/LOCODE
Name	Name of the location, with diacritics
NameWoDiacritics	The same name with diacritics replaced by ASCII letters
Subdivision	ISO 3166-2 subdivision code, for example DU. Defined in subdivision-codes.csv. May be empty
Status	Approval or verification status, for example AI. Defined in status-indicators.csv
Function	8-character string, one position for each function (port, rail terminal, airport and so on). A digit or letter means the function is present, a dash means it is not. Defined in function-classifiers.csv
Date	Date of the entry as YYMM, for example 2501 is January 2025. May be empty
IATA	IATA code, only when it differs from the UN/LOCODE location code. May be empty
Coordinates	Latitude and longitude as degrees and minutes with a letter for the direction, for example 2515N 05516E. May be empty
Remarks	Free-text remark. May be empty

Known limitations

• The file follows UNECE's "Pre-Release (Latest)", not an official release. UNECE describes it as "for latest codes in between official releases", so it can contain codes approved after the last official biannual publication, and entries can still change.
• Country + Location is not unique. Some places have two entries, for example an old and a changed name (one with # in Change). Use Change and Date to tell them apart.
• Rows with Change set to X are entries UNECE marks for removal. They stay in the file until a later release drops them.
• The code for Namibia is the two letters NA (in Country and in SUCountry). Many tools read the text NA as "missing". Read the files as text and keep empty cells empty: pd.read_csv("data/code-list.csv", dtype=str, keep_default_na=False).
• Codes and Function strings are text. Location can be made only of digits and letters such as 5MS, and Function starts with digits or dashes. Do not let your tool turn them into numbers.
• Coordinates is one text column in degrees and minutes, not two decimal numbers.
• Rows without a location code (country-name rows and alias rows) are not in code-list.csv. The aliases are in alias.csv.

Data sources

• UN/LOCODE, published by UNECE: the code list and its lookup tables. The build uses the file UNECE calls "Pre-Release (Latest)". Licence: unclear (see License).

Changes made to the source data: the script reads the release's Access database for the lookup tables and its three CSV parts for the code list. For code-list.csv it joins the three parts, moves the alias rows (Change set to =) to alias.csv, drops the country-name rows, swaps Status and Function where the release has them the wrong way round, removes duplicate rows (same country, location, name, subdivision, date and coordinates), and writes the columns in a fixed order. For subdivision-codes.csv it adds the subdivision type from the release's CSV file, collapses extra spaces and removes repeated lines. All files are written with LF line endings.

Not endorsed by, or affiliated with, UNECE or the United Nations.

Preparation

You need Python 3.12 or greater, make, and the tools mdbtools and gawk. Install them (macOS shown; on Ubuntu use sudo apt-get install mdbtools gawk):

brew install mdbtools gawk
pip install -r scripts/requirements.txt csvkit

Download the release into release/ and build the data:

./scripts/fetch_release.sh
make

scripts/fetch_release.sh downloads https://unlocode.unece.org/downloads/unlocode-latest.zip. To build from another zip, for example an official release, run UNLOCODE_URL=<zip url> ./scripts/fetch_release.sh. The zip must contain this folder layout:

release/
  UNLOCODE CodeList.mdb
  csv/
    SubdivisionCodes.csv
    UNLOCODE CodeListPart1.csv
    UNLOCODE CodeListPart2.csv
    UNLOCODE CodeListPart3.csv

Set PYTHON=<path> if python3 is not the interpreter that has the requirements installed (make PYTHON=.venv/bin/python).

A GitHub workflow (.github/workflows/update.yml) runs this on the first day of every month (at 03:00 UTC) and when started by hand. It commits data/ if something changed.

License

The licence of the UN/LOCODE data is unclear, because UNECE's own pages say different things. What we checked:

• The terms of the UN/LOCODE site say the United Nations grants permission "to download and copy the information, documents and materials ... for the User's personal, non-commercial use, without any right to resell or redistribute them or to compile or create derivative works therefrom", and that this is "subject to ... more specific restrictions that may apply to specific Material".
• The footer of the same site says: "All UN/CEFACT standards are free to use under CC By 4.0 license". UN/LOCODE is published through UN/CEFACT (Recommendation 16). The pages we read do not say whether that sentence covers the code list itself.
• Recommendation 16 says updated versions of the directory "will be published on the UNECE website free of charge".
• The source repository has a GPL-3.0 LICENSE file for the repository. Its data folder has no licence notice of its own.

We could not establish which of these applies to the data in this package. If you plan to redistribute it or use it commercially, ask UNECE for written confirmation first. We credit UNECE as the source in any case (see ATTRIBUTION.txt).

The maintainers place only their own work (the scripts, the column definitions and the file layout) under the Public Domain Dedication and License (PDDL). This does not cover the UN/LOCODE data itself.