The schema docs say:
But there are many records where this isn't the case.
Sample Query:
duckdb -csv -header -cmd "LOAD httpfs" -cmd "SET s3_region='us-west-2'" -c "SELECT id, division_id, country, region, subtype, class, is_land, is_territorial, version, struct_extract(names, 'primary') AS name_primary, sources FROM read_parquet('s3://overturemaps-us-west-2/release/2026-06-17.0/theme=divisions/type=division_area/part-00006-70bf6111-3c8d-5cd6-b586-4a8e4852ee57-c000.zstd.parquet') WHERE id = '59e53dd7-e5f7-482c-bcba-c91abb74f7da'"Results in
id: 59e53dd7-e5f7-482c-bcba-c91abb74f7da
division_id: a3f5d985-9d91-4629-a530-f66420a0be01
name: 灣仔區 Wan Chai District
subtype: region
class: land
is_land: true
is_territorial: true
version: 4
source: OpenStreetMap r2558883@20
Sampling Hong Kong data, it looks like divisions which have a consistent (land vs land + nautical) separation follow the is_land and is_territory schema constraints, but for divisions where this distinction isn't made, they are both set to true.
For reference, 灣仔區 Wan Chai District is not land-locked, and could be separated into land vs nautical divisions.
The schema docs say:
But there are many records where this isn't the case.
Sample Query:
Results in
Sampling Hong Kong data, it looks like divisions which have a consistent (land vs land + nautical) separation follow the
is_landandis_territoryschema constraints, but for divisions where this distinction isn't made, they are both set to true.For reference, 灣仔區 Wan Chai District is not land-locked, and could be separated into land vs nautical divisions.