zarr-python - Work with chunked Zarr arrays in Python
Create, read, write, compress, resize, and store chunked N-dimensional arrays locally, in memory, in ZIP files, or through cloud backends.
Tags
Updated: 2026-10-15Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Create chunked arrays
- Read and write slices
- Resize and append arrays
- Configure compression codecs
- Use local memory ZIP stores
- Access S3 and GCS stores
- Integrate NumPy Dask Xarray
Inputs
- Array shape and dtype
- Chunk and shard shapes
- Input array data
- Store path or URI
- Access patterns
- Compression settings
- Cloud storage options
- Provider credentials
Outputs
- Zarr arrays
- NumPy array data
- Local or cloud stores
- Compressed chunk data
- Resized array state
- Appended array data
Requirements
- Python 3.12 or newer
- Zarr 3.x
- NumPy 2.0 or newer
- zarr[remote] for remote I/O
- Pinned s3fs or gcsfs backends
- Cloud provider authentication