Zarr Encoding Specification#

Dimension Encoding in Zarr Formats#

Xarray encodes array dimensions differently depending on the Zarr format version:

Zarr V2 Format: Xarray uses a special Zarr array attribute: _ARRAY_DIMENSIONS. The value of this attribute is a list of dimension names (strings), for example ["time", "lon", "lat"]. When writing data to Zarr V2, Xarray sets this attribute on all variables based on the variable dimensions. This attribute is visible when accessing arrays directly with zarr-python.

Zarr V3 Format: Xarray uses the native dimension_names field in the array metadata. This is part of the official Zarr V3 specification and is not stored as a regular attribute. When accessing arrays with zarr-python, this information is available in the array’s metadata but not in the attributes dictionary.

When reading a Zarr group, Xarray looks for dimension information in the appropriate location based on the format version, raising an error if it can’t be found. The dimension information is used to define the variable dimension names and then (for Zarr V2) removed from the attributes dictionary returned to the user.

CF Conventions#

Xarray uses its standard CF encoding/decoding functionality for handling metadata (see decode_cf()). This includes encoding concepts such as dimensions and coordinates. The coordinates attribute, which lists coordinate variables (e.g., "yc xc" for spatial coordinates), is one part of the broader CF conventions used to describe metadata in NetCDF and Zarr.

Compatibility and Reading#

Because of these encoding choices, Xarray cannot read arbitrary Zarr arrays, but only Zarr data with valid dimension metadata. Xarray supports:

• Zarr V2 arrays with _ARRAY_DIMENSIONS attributes

• Zarr V3 arrays with dimension_names metadata

• NCZarr format (dimension names are defined in the .zarray file)

After decoding the dimension information and assigning the variable dimensions, Xarray proceeds to [optionally] decode each variable using its standard CF decoding machinery used for NetCDF data.

Finally, it’s worth noting that Xarray writes (and attempts to read) “consolidated metadata” by default (the .zmetadata file), which is another non-standard Zarr extension, albeit one implemented upstream in Zarr-Python. You do not need to write consolidated metadata to make Zarr stores readable in Xarray, but because Xarray can open these stores much faster, users will see a warning about poor performance when reading non-consolidated stores unless they explicitly set consolidated=False. See Consolidated Metadata for more details.

Examples: Zarr Format Differences#

The following examples demonstrate how dimension and coordinate encoding differs between Zarr format versions. We’ll use the same tutorial dataset but write it in different formats to show what users will see when accessing the files directly with zarr-python.

Example 1: Zarr V2 Format

import os
import xarray as xr
import zarr
# Load tutorial dataset and write as Zarr V2
ds = xr.tutorial.load_dataset("rasm")
ds.to_zarr("rasm_v2.zarr", mode="w", consolidated=False, zarr_format=2)
# Open with zarr-python and examine attributes
zgroup = zarr.open("rasm_v2.zarr")
print("Zarr V2 - Tair attributes:")
tair_attrs = dict(zgroup["Tair"].attrs)
for key, value in tair_attrs.items():
    print(f"  '{key}': {repr(value)}")

Zarr V2 - Tair attributes:
  'units': 'C'
  'long_name': 'Surface air temperature'
  'type_preferred': 'double'
  'time_rep': 'instantaneous'
  'coordinates': 'yc xc'
  '_ARRAY_DIMENSIONS': ['time', 'y', 'x']

Example 2: Zarr V3 Format

# Write the same dataset as Zarr V3
ds.to_zarr("rasm_v3.zarr", mode="w", consolidated=False, zarr_format=3)
# Open with zarr-python and examine attributes
zgroup = zarr.open("rasm_v3.zarr")
print("Zarr V3 - Tair attributes:")
tair_attrs = dict(zgroup["Tair"].attrs)
for key, value in tair_attrs.items():
    print(f"  '{key}': {repr(value)}")
# For Zarr V3, dimension information is in metadata
tair_array = zgroup["Tair"]
print(f"\nZarr V3 - dimension_names in metadata: {tair_array.metadata.dimension_names}")

Zarr V3 - Tair attributes:
  'units': 'C'
  'long_name': 'Surface air temperature'
  'type_preferred': 'double'
  'time_rep': 'instantaneous'
  'coordinates': 'yc xc'
  '_FillValue': 'AAAAAAAAnkc='
Zarr V3 - dimension_names in metadata: ('time', 'y', 'x')

Chunk Key Encoding#

When writing data to Zarr stores, Xarray supports customizing how chunk keys are encoded through the chunk_key_encoding parameter in the variable’s encoding dictionary. This is particularly useful when working with Zarr V2 arrays and you need to control the dimension separator in chunk keys.

For example, to specify a custom separator for chunk keys:

import xarray as xr
import numpy as np
from zarr.core.chunk_key_encodings import V2ChunkKeyEncoding
# Create a custom chunk key encoding with "/" as separator
enc = V2ChunkKeyEncoding(separator="/").to_dict()
# Create and write a dataset with custom chunk key encoding
arr = np.ones((42, 100))
ds = xr.DataArray(arr, name="var1").to_dataset()
ds.to_zarr(
    "example.zarr",
    zarr_format=2,
    mode="w",
    encoding={"var1": {"chunks": (42, 50), "chunk_key_encoding": enc}},
)

<xarray.backends.zarr.ZarrStore at 0x7987c418efc0>

The chunk_key_encoding option accepts a dictionary that specifies the encoding configuration. For Zarr V2 arrays, you can use the V2ChunkKeyEncoding class from zarr.core.chunk_key_encodings to generate this configuration. This is particularly useful when you need to ensure compatibility with specific Zarr V2 storage layouts or when working with tools that expect a particular chunk key format.