The whole dataset in one format. Cached one year. Byte ranges are offered on Parquet, DuckDB, SQLite, Arrow, Excel, GeoPackage, PMTiles and csv.gz. CSV, NDJSON, JSON and GeoJSON are sent whole, gzipped when the client accepts it, so a lazy or range reader should use parquet or csv.gz.
Request
- Base URL: https://publicdata.au
- URL: https://publicdata.au/d/{slug}/v/{version}/data.{format}
- Auth: none declared
Path parameters
The dataset's slug, as in its page URL /d/<slug>/.
A dated version from versions.json.
parquet, csv, csv.gz, ndjson and duckdb are on every table version. xlsx is there up to 50 MB of CSV and 1,048,575 rows, json up to 50 MB of NDJSON, and sqlite up to 500 MB of CSV. gpkg, and geo.parquet for points, exist only for datasets that declare geometry, with geojson up to a 100 MB file, and pmtiles only for polygon and line layers. A version's manifest.json lists each format the limits left out, with the reason, under formats_left_out, and that set never changes. A version whose manifest has no caps field was fetched before these limits: it also has arrow, its xlsx stops at the Excel row limit, and its json and geojson stop at 2,000,000 rows.
Response
The file
Changes
No recorded changes to this operation across all 1 revision of this API.