zarr-python - Create chunked arrays for cloud workflows
Store and process large N-dimensional arrays with chunking, compression, parallel I/O, cloud storage, and NumPy, Dask, or Xarray integration.
Tags
Updated: 2026-10-06Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Create chunked arrays
- Read and write slices
- Resize and append arrays
- Configure compression codecs
- Use local and cloud stores
- Organize arrays in groups
- Attach array metadata
- Tune chunk layouts
- Shard large arrays
Inputs
- Array shape
- Chunk shape
- Data type
- Input array data
- Storage path or backend
- Compression settings
- Cloud storage credentials
- Access pattern
Outputs
- Persisted Zarr arrays
- Read NumPy arrays
- Resized or appended arrays
- Compressed storage objects
- Array and group metadata
- Storage hierarchy
Requirements
- Python 3.11 or newer
- zarr package
- s3fs for S3 storage
- gcsfs for Google Cloud Storage
- Cloud storage credentials for authenticated access