Skip to main content

Converting & exporting to Zarr

Eider writes Zarr two ways, both via the CLI:

  • eider ingest — convert a legacy file (NetCDF, GeoTIFF, CSV with geometry) to a GeoZarr array.
  • eider export — write the result of a DuckDB query to a Zarr array.

Convert a legacy file​

eider ingest input.geojson out.zarr --value-column value

Override automatic chunking with --chunks, e.g. --chunks '{"time": 30}'. See eider ingest.

Export a query result​

eider export runs a SQL query and writes the --value-column as the array values; every other column is treated as a 0-based integer coordinate index.

eider export \
  --db analysis.duckdb \
  --query "SELECT t, y, x, value FROM gridded" \
  --dest out.zarr \
  --value-column value

The coordinate columns (t, y, x above) must be 0-based integer dimension indices; the array shape is inferred from their distinct counts. See eider export.

Next steps​