More robust dtype check

main branch of zarr is not a higher version right now.

Author: ianhi

xarray/backends/zarr.py

@@ -9,7 +9,6 @@
 
 import numpy as np
 import pandas as pd
-from packaging.version import Version
 
 from xarray import coding, conventions
 from xarray.backends.chunks import grid_rechunk, validate_grid_chunks_alignment
@@ -819,9 +818,9 @@ def open_store_variable(self, name):
         elif "_FillValue" in attributes:
             # TODO update version check for the released version with dtypes
             #  probably be 3.1
-            import zarr
 
-            if Version(zarr.__version__) > Version("3.0.9"):
+            # if Version(zarr.__version__) > Version("3.0.10"):
+            if hasattr(zarr_array.metadata.data_type, "to_native_dtype"):
                 native_dtype = zarr_array.metadata.data_type.to_native_dtype()
                 attributes["_FillValue"] = (
                     # Use the new dtype infrastructure instead of doing xarray
@@ -961,12 +960,11 @@ def store(
             # To do so, we decode variables directly to access the proper encoding,
             # without going via xarray.Dataset to avoid needing to load
             # index variables into memory.
-            store_vars = {
-                k: self.open_store_variable(name=k) for k in existing_variable_names
-            }
 
             existing_vars, _, _ = conventions.decode_cf_variables(
-                variables=store_vars,
+                variables={
+                    k: self.open_store_variable(name=k) for k in existing_variable_names
+                },
                 # attributes = {} since we don't care about parsing the global
                 # "coordinates" attribute
                 attributes={},

xarray/tests/__init__.py

@@ -130,20 +130,26 @@ def _importorskip(
 has_rasterio, requires_rasterio = _importorskip("rasterio")
 has_zarr, requires_zarr = _importorskip("zarr")
 has_zarr_v3, requires_zarr_v3 = _importorskip("zarr", "3.0.0")
-has_zarr_v3_dtypes, requires_zarr_v3_dtypes = _importorskip("zarr", "3.0.9")
-
-# Additional check for zarr dtype support (dev versions > 3.0.9)
-if has_zarr:
+has_zarr_v3_dtypes, requires_zarr_v3_dtypes = _importorskip("zarr", "3.0.10")
+if has_zarr_v3:
     import zarr
 
-    # Dev versions like "3.0.9.dev47+g9da38f75" are > "3.0.9"
-    has_zarr_v3_dtypes = has_zarr_v3_dtypes or (
-        Version(zarr.__version__) > Version("3.0.9")
-    )
+    has_zarr_v3_dtypes = hasattr(zarr.core, "dtype")
+
+    # has_zarr_v3_dtypes = hasattr(zarr.Array.metadata.data_type, "to_native_dtype")
     requires_zarr_v3_dtypes = pytest.mark.skipif(
-        not has_zarr_v3_dtypes, reason="requires zarr>3.0.9 (including dev versions)"
+        not has_zarr_v3_dtypes, reason="requires zarr>3.0.10 (including dev versions)"
     )
 
+# Additional check for zarr dtype support (dev versions > 3.0.9)
+# if has_zarr:
+#     import zarr
+#
+#     # Dev versions like "3.0.10.dev47+g9da38f75" are > "3.0.10"
+#     has_zarr_v3_dtypes = has_zarr_v3_dtypes or (
+#         Version(zarr.__version__) > Version("3.0.10")
+#     )
+
 has_fsspec, requires_fsspec = _importorskip("fsspec")
 has_iris, requires_iris = _importorskip("iris")
 has_numbagg, requires_numbagg = _importorskip("numbagg")

xarray/tests/test_backends.py

@@ -2908,8 +2908,11 @@ def test_append_with_existing_encoding_raises(self) -> None:
 
     @pytest.mark.parametrize("dtype", ["U", "S"])
     def test_append_string_length_mismatch_raises(self, dtype) -> None:
-        if has_zarr_v3() and not has_zarr_v3_dtypes:
-            pytest.skip("This works on pre dtype updated zarr python")
+        print("start")
+        print(has_zarr_v3)
+        print(has_zarr_v3_dtypes)
+        if has_zarr_v3 and not has_zarr_v3_dtypes:
+            skip_if_zarr_format_3("This actually works fine with Zarr format 3")
 
         ds, ds_to_append = create_append_string_length_mismatch_test_data(dtype)
         with self.create_zarr_target() as store_target:
@@ -2919,6 +2922,22 @@ def test_append_string_length_mismatch_raises(self, dtype) -> None:
                     store_target, append_dim="time", **self.version_kwargs
                 )
 
+    @pytest.mark.parametrize("dtype", ["U", "S"])
+    def test_append_string_length_mismatch_works(self, dtype) -> None:
+        skip_if_zarr_format_2("This doesn't work with Zarr format 2")
+        # ...but it probably would if we used object dtype
+        if has_zarr_v3_dtypes:
+            pytest.skip("This works on pre dtype updated zarr python")
+
+        ds, ds_to_append = create_append_string_length_mismatch_test_data(dtype)
+        expected = xr.concat([ds, ds_to_append], dim="time")
+
+        with self.create_zarr_target() as store_target:
+            ds.to_zarr(store_target, mode="w", **self.version_kwargs)
+            ds_to_append.to_zarr(store_target, append_dim="time", **self.version_kwargs)
+            actual = xr.open_dataset(store_target, engine="zarr")
+            xr.testing.assert_identical(expected, actual)
+
     def test_check_encoding_is_consistent_after_append(self) -> None:
         ds, ds_to_append, _ = create_append_test_data()