Part of a Data Solution

Postal Codes Solution
Every worldwide postal-code dataset in one bundle — the ultimate global reference for precise postal codes.
Explore →

Download · CSV · 103 kB Explore with AI Report an Issue

API Access

Access dataset files directly from scripts, code, or AI agents.

Browse dataset files

Dataset Files

Each file has a stable URL (r-link) that you can use directly in scripts, apps, or AI agents. These URLs are permanent and safe to hardcode.

/logistics/postal-codes-gr/

├ .licensing-notes.md

https://datahub.io/logistics/postal-codes-gr/_r/-/.licensing-notes.md

├ .migration-notes.md

https://datahub.io/logistics/postal-codes-gr/_r/-/.migration-notes.md

├ ATTRIBUTION.txt

https://datahub.io/logistics/postal-codes-gr/_r/-/ATTRIBUTION.txt

├ README.md

https://datahub.io/logistics/postal-codes-gr/_r/-/README.md

└ datapackage.yaml

https://datahub.io/logistics/postal-codes-gr/_r/-/datapackage.yaml

Key Files

Start with these files — they give you everything you need to understand and access the dataset.

datapackage.yaml— metadata & schema

https://datahub.io/logistics/postal-codes-gr/_r/-/datapackage.yaml

README.md— documentation

https://datahub.io/logistics/postal-codes-gr/_r/-/README.md

Typical Usage

1. Fetch datapackage.yaml to inspect schema and resources
2. Download data resources listed in datapackage.yaml
3. Read README.md for full context

Data Files

Explore with AI

postal-codes-gr

Field	Type	Description	Constraints	Title
country_code	string	ISO 3166-1 alpha-2 country code -- always GR.		Country Code
postal_code	string	5-digit Greek postal code (GISCO POSTCODE). Unique; the producer checks it.	{ "pattern": "^\\d{5}$" }	Postal Code
place_name	string	Community or pseudo-community name as GISCO spells it (LAU_NAME, without Greek accents). A 'Ψευδοκοινοτητα X' (pseudo-community) label stands for a municipality with no community subdivisions.		Place Name
admin_name1	string	Region (Περιφέρεια) -- 13 regions, derived from the NUTS2 part of the NUTS3 code.		Administrative Name 1
admin_code1	string	ISO 3166-2:GR region letter.		Administrative Code 1
admin_name2	string	Regional unit(s) (Περιφερειακή Ενότητα) of the NUTS3 area -- 52 NUTS3 areas.		Administrative Name 2
admin_code2	string	NUTS 2024 level-3 code.		Administrative Code 2
admin_name3	string	Not populated -- municipality (Δήμος) name not resolved in this release.		Administrative Name 3
admin_code3	string	ELSTAT NSI 2021 municipality code as GISCO gives it (some have lost their leading zero); name not resolved.		Administrative Code 3
latitude	number	Not populated -- the GISCO CSV distribution carries no geometry.		Latitude
longitude	number	Not populated -- the GISCO CSV distribution carries no geometry.		Longitude
accuracy	string	Not populated -- no coordinates in this release.		Accuracy
alternative_city_name	string	Not populated in this release.		Alternative City Name

Download

Download CSV

About

Last updated: 5 October 2026
Total rows: 1,041
Format: CSV
File size: 103 kB
Source: Eurostat GISCO postal code point dataset 2025
License: Creative Commons Attribution-ShareAlike 4.0

About this dataset

Greece (GR): migration notes

Internal record of the P6 migration from the old scripts to the producer: the discrepancies found, their causes, and who decided what. This is a dated record, so the numbers are as of the migration and not kept current. It is not shipped in the zip.

Issue: pc-28e.8.9
Commit: worktree commit on top of c51b72a (see the commit P6 gr: SST producer; …)
Parity verdict: format-only (gate: pass, exit 0) against the live public CSV
Producer: custom, refresh disabled

Sources

Input	URL	Notes
Eurostat GISCO postal code point dataset 2025, CSV distribution	https://gisco-services.ec.europa.eu/distribution/v2/pcode/csv/PCODE_PT_2025_4326.csv	All countries (~136 MB); the producer keeps `CNTR_ID == 'EL'`. CC BY-SA 4.0

The old build_postal_codes_product.py read a local sources/gisco_gr_filtered.csv. Its docstring names this URL as where that snapshot came from, so the producer fetches it.
The URL is year-versioned (PCODE_PT_2025_…). It stays pinned to 2025, the release live was built from. A newer GISCO year is a new file name, so moving to it is a deliberate change, not a refresh. Fetched 2026-10-05: 136,116,955 bytes, SHA-256 459efc82…, the same file as on 2026-10-02.
Python's default CA store fails GISCO's TLS chain on this machine, so the producer fetches with requests (certifi), as es does.

Parity against the live public CSV (2026-10-05)

Rows: 1,041 live and 1,041 fresh, none only on one side, 0 changed cells. The only difference is line endings: live is CRLF (the old csv.DictWriter default), the producer writes LF. With \r stripped, the fresh CSV is byte-identical to live. Covered by the standing sign-off (CRLF→LF only, equal rows, 0 changed cells).

Decisions

Decision	By	Basis
Reconstruct the filter + NUTS lookup step inside the producer	Agent; user accepted as a port (2026-10-05)	The old builder's input was already filtered and enriched, and the step that did it was never committed. Its docstring describes the step ("derived here from NUTS3 code via a hardcoded, verified lookup"), and the source is the one it names, so this is treated as a port, not a new pipeline. The 13-region and 52-NUTS3 name tables were read off the live CSV (every code maps to exactly one name there) and equal the earlier, discarded reconstruction in `2b7224e`. Parity confirms them: 0 changed cells.
Keep the two lookup tables inline in `produce.py`	Agent: port choice	Small, stable classification; cz keeps its crosswalk inline the same way
Keep NSI_2021 raw in `admin_code3`, including 7-digit codes where GISCO dropped a leading zero	Agent: a faithful port	The old builder kept it raw; fixing it would change live data
Write LF line endings	Standing rule (CRLF→LF pre-approved)	Same as every other producer
Old exact row-count assert dropped; structural checks kept as `ProducerError`	Brief	Unique 5-digit postal codes, known NUTS3, all 13 regions, non-empty LAU_NAME, 7–8 digit NSI_2021
Types: `latitude`/`longitude` `number`; `accuracy` stays `string`; codes stay `string`; `primaryKey: [postal_code]`	Agent, from the produced CSV	All coordinate/accuracy cells are empty (GISCO's CSV has no geometry); unpopulated `accuracy` is `string` elsewhere too (au, nz). `postal_code` is unique and the producer enforces it
Counts removed from the descriptor description and README (row count, coverage %, the 27% pseudo-community share)	Brief (doc rules)	Region (13) and NUTS3 (52) counts kept as admin-unit counts

Transforms carried over

Keep CNTR_ID == 'EL'; postal_code = POSTCODE; place_name = LAU_NAME as GISCO spells it (no Greek accents).
admin_code2 = NUTS3_2024 with GISCO's literal quotes stripped; admin_name2 = English regional-unit name from the NUTS3 table.
admin_code1/admin_name1 = ISO 3166-2:GR letter and English region name from the NUTS2 prefix.
admin_code3 = raw NSI_2021; admin_name3, coordinates, accuracy and alternative_city_name empty.
Rows sorted by postal_code.

Not ported

The old builder's packaging (datapackage.json, README, ATTRIBUTION, zip) and publish_r2.py: build.py and the shared publisher own those now.

Open questions

Resolved (user, 2026-10-05): reconstructing the never-committed filter/lookup step counts as a port.
Resolved (orchestrator, 2026-10-05): removed the false "sample of the first 100 rows… contact sales" note from README.md (gr is free and its full CSV is public).
.licensing-notes.md and ATTRIBUTION-style text in it still carry historical counts (1,041 rows, 83%); left for the pc-hfz sweep.
148 admin_code3 values are 7 digits because GISCO dropped a leading zero (e.g. 7010100 vs GISCO_2021 EL_07010100). Padding to 8 digits would be a data change; not done.