Zarr.jlA Julia library for the Zarr storage format

Chunked, compressed N-dimensional arrays for every scale and storage backend

Zarr v3 & v2 Support

Compatible with both Zarr v2 and Zarr v3 formats, with examples provided for each version wherever possible to illustrate equivalent usage and highlight any differences between them.

Chunked Arrays & Codecs

Slice through terabytes effortlessly, tune chunk shapes and pick from a rich set of compression codecs to match your access patterns and storage constraints.

Flexible Storage Backends

Read and write to local disk or in-memory stores. S3-compatible object storage is supported for reading. New backends can be added by subtyping Zarr.AbstractStore.

Groups & Hierarchies

Structure your data the way you think about it, nest arrays into groups and build intuitive hierarchies for any dataset complexity.

YAXArrays.jl Integration

Zarr.jl integrates seamlessly with YAXArrays.jl, providing the storage layer for labeled, multi-dimensional data.

Sharding - Coming Soon!

Support for the Zarr v3 sharding codec is on the roadmap, enabling efficient storage of many chunks within a single file.