Skip to documentation content
Dataset APIDocs

Downloading releases

Stream complete release artifacts instead of stretching bounded queries.

In this guide
  1. 1Choose a format
  2. 2Stream to storage
  3. 3Preserve the manifest
In this guide
  1. 1Choose a format
  2. 2Stream to storage
  3. 3Preserve the manifest

Use downloads for complete ingestion

Use a query for bounded interactive reads and a download when the job must ingest the complete release. Downloads are addressed by dataset and release IDs, preserving the same reproducibility model.

Choose a declared format

formatResponse typeTypical use
jsonlapplication/x-ndjsonStreaming and line-oriented pipelines
csvtext/csvTabular tools and interoperable exports
parquetapplication/octet-streamColumnar analytics workflows

Stream instead of buffering

cURL
curl --fail --location \
  "https://api.datasets.nordicdevhouse.com/v1/datasets/finnish-vehicle-market-listings/releases/dsv-preview-finnish-vehicle-market-listings/download?format=jsonl" \
  -H "Authorization: Bearer $DATASET_API_KEY" \
  -H "X-API-Version: 1" \
  --output release.jsonl

Write to a temporary location, validate that the transfer completed, then atomically promote the file into downstream processing. Avoid loading an entire release into memory.

Archive provenance with the artifact

Was this page helpful?