Try to find postal codes without boundaries in OSM.

output.txt

["01170", "01290", "01410", "01750", "06000", "06100", "06130", "06150", "06160", "06200", "06300", "06400", "06520", "06600", 
"07160", "07190", "07310", "07320", "07510", "07630", "12140", "12350", "12460", "13090", "13100", "20000", "20090", "20200", 
"20220", "20251", "20256", "20270", "20600", "26340", "26470", "30000", "30900", "33950", "33970", "34000", "34070", "34080", 
"34090", "35000", "35200", "35700", "37000", "37100", "37200", "38000", "39120", "39330", "39600", "39800", "42000", "42100", 
"44000", "44100", "44200", "44300", "45000", "45100", "46630", "49000", "49100", "50100", "50130", "57000", "57050", "57070", 
"59000", "59260", "59491", "59493", "59650", "59777", "59800", "60113", "60190", "63000", "63100", "65200", "65710", "66000", 
"66100", "67000", "67100", "67130", "67200", "67570", "68126", "68630", "75000", "76450", "76540", "76600", "76610", "76620", 
"83000", "83100", "83200", "83370", "83530", "83600", "83700", "84000", "84140", "87000", "87100", "87280", "92190", "92360"]

postal_codes.py

#!/usr/bin/env python
import json
import logging
import os

from lxml import etree

logging_level = logging.WARN
if 'DEBUG' in os.environ:
    logging_level = logging.DEBUG
logging.basicConfig(level=logging_level)

# communes.xml file is generated with:
#   osmosis \
#       --read-pbf france-latest.osm.pbf \
#       --tf accept-relations boundary=administrative \
#       --tf accept-relations admin_level=8
#       --tf reject-ways \
#       --tf reject-nodes \
#       --write-xml communes.xml
with open('communes.xml', 'r') as fh:
    communes_tree = etree.parse(fh)

# Extract all multiple postal codes fields from the "communes.xml" file
# For these cases, the city boundary does not match with the postal code.
multiple_postcodes_from_communes = []
for relation in communes_tree.findall('relation'):
    relation_id = relation.attrib['id']

    postcode = None
    for tag in relation.findall('tag'):
        if tag.attrib['k'] in ['addr:postcode', 'postal_code']:
            postcode = [x.strip() for x in tag.attrib['v'].split(';')]
            break

    if not postcode:
        continue

    if len(postcode) > 1:
        multiple_postcodes_from_communes.extend(postcode)
        logging.info(
            '[communes] Multiple postcodes found for relation %s: %s.',
            relation_id, ','.join(postcode)
        )

# "postal_codes.xml" file is generated with:
#   osmosis \
#       --read-pbf france-latest.osm.pbf \
#       --tf accept-relations boundary=postal_code \
#       --tf reject-ways \
#       --tf reject-nodes \
#       --write-xml postal_codes.xml
with open('postal_codes.xml', 'r') as fh:
    postal_codes_tree = etree.parse(fh)

# Extract all the specific boundaries for postal codes (from
# boundary=postal_code relations).
postal_codes_with_boundary = []
for relation in postal_codes_tree.findall('relation'):
    relation_id = relation.attrib['id']
    postcode = None
    for tag in relation.findall('tag'):
        if tag.attrib['k'] == 'postal_code':
            postcode = tag.attrib['v']
            break
    postal_codes_with_boundary.append(postcode)
    logging.info(
        '[postal_codes] Found boundary for postal code %s.',
        postcode
    )


# "arrondissements.xml" file is generated with:
#   osmosis \
#       --read-pbf france-latest.osm.pbf \
#       --tf accept-relations boundary=administrative \
#       --tf accept-relations admin_level=9
#       --tf reject-ways \
#       --tf reject-nodes \
#       --write-xml arrondissements.xml
with open('arrondissements.xml', 'r') as fh:
    arrondissements_tree = etree.parse(fh)

multiple_postcodes_from_arrondissements = []
# Extract all the arrondissements with a unique postal code.
for relation in arrondissements_tree.findall('relation'):
    relation_id = relation.attrib['id']
    postcode = None
    for tag in relation.findall('tag'):
        if tag.attrib['k'] in ['addr:postcode', 'postal_code']:
            postcode = [x.strip() for x in tag.attrib['v'].split(';')]
            break
    if not postcode:
        continue
    if len(postcode) > 1:
        multiple_postcodes_from_arrondissements.extend(postcode)
    else:
        postal_codes_with_boundary.extend(postcode)
    logging.info(
        '[postal_codes] Found arrondissement for postal code %s.',
        postcode
    )


# We can now list the postal codes without any matching boundary.
# NOTE: Considering further levels of admin_level such as admin_level=10 does
# not reduce further this list.
postal_codes_without_boundaries = sorted(list(
    set(
        multiple_postcodes_from_communes +
        multiple_postcodes_from_arrondissements
    ) - set(postal_codes_with_boundary)
))
# Filter out postal codes with less than 5 characters which are not French
# ones
print(json.dumps([x for x in postal_codes_without_boundaries if len(x) == 5]))