align_chunks not working for datasets

doc/whats-new.rst

@@ -25,6 +25,9 @@ Deprecations
 Bug fixes
 ~~~~~~~~~
 
+- Fix the ``align_chunks`` parameter on the :py:meth:`~xarray.Dataset.to_zarr` method, it was not being
+  passed to the underlying :py:meth:`~xarray.backends.api` method (:issue:`10501`, :pull:`10516`).
+
 
 Documentation
 ~~~~~~~~~~~~~

xarray/backends/chunks.py

@@ -138,9 +138,6 @@ def build_grid_chunks(
     chunk_size: int,
     region: slice | None = None,
 ) -> tuple[int, ...]:
-    if region is None:
-        region = slice(0, size)
-
     region_start = region.start or 0
     # Generate the zarr chunks inside the region of this dim
     chunks_on_region = [chunk_size - (region_start % chunk_size)]
@@ -159,6 +156,13 @@ def grid_rechunk(
     if not nd_var_chunks:
         return v
 
+    # This is useful for the scenarios where the enc_chunks are bigger than the
+    # variable chunks, which happens when the user specifies the enc_chunks manually.
+    enc_chunks = tuple(
+        min(enc_chunk, sum(v_chunk))
+        for enc_chunk, v_chunk in zip(enc_chunks, v.chunks, strict=True)
+    )
+
     nd_grid_chunks = tuple(
         build_grid_chunks(
             sum(var_chunks),
@@ -191,9 +195,9 @@ def validate_grid_chunks_alignment(
     base_error = (
         "Specified Zarr chunks encoding['chunks']={enc_chunks!r} for "
         "variable named {name!r} would overlap multiple Dask chunks. "
-        "Check the chunk at position {var_chunk_pos}, which has a size of "
-        "{var_chunk_size} on dimension {dim_i}. It is unaligned with "
-        "backend chunks of size {chunk_size} in region {region}. "
+        "Please check the Dask chunks at position {var_chunk_pos} and "
+        "{var_chunk_pos_next}, on axis {axis}, they are overlapped "
+        "on the same Zarr chunk in the region {region}. "
         "Writing this array in parallel with Dask could lead to corrupted data. "
         "To resolve this issue, consider one of the following options: "
         "- Rechunk the array using `chunk()`. "
@@ -202,7 +206,7 @@ def validate_grid_chunks_alignment(
         "- Enable automatic chunks alignment with `align_chunks=True`."
     )
 
-    for dim_i, chunk_size, var_chunks, interval, size in zip(
+    for axis, chunk_size, var_chunks, interval, size in zip(
         range(len(enc_chunks)),
         enc_chunks,
         nd_var_chunks,
@@ -215,9 +219,10 @@ def validate_grid_chunks_alignment(
                 raise ValueError(
                     base_error.format(
                         var_chunk_pos=i + 1,
+                        var_chunk_pos_next=i + 2,
                         var_chunk_size=chunk,
+                        axis=axis,
                         name=name,
-                        dim_i=dim_i,
                         chunk_size=chunk_size,
                         region=interval,
                         enc_chunks=enc_chunks,
@@ -237,9 +242,10 @@ def validate_grid_chunks_alignment(
                 raise ValueError(
                     base_error.format(
                         var_chunk_pos=0,
+                        var_chunk_pos_next=0,
                         var_chunk_size=var_chunks[0],
+                        axis=axis,
                         name=name,
-                        dim_i=dim_i,
                         chunk_size=chunk_size,
                         region=interval,
                         enc_chunks=enc_chunks,
@@ -251,9 +257,10 @@ def validate_grid_chunks_alignment(
 
             error_on_last_chunk = base_error.format(
                 var_chunk_pos=len(var_chunks) - 1,
+                var_chunk_pos_next=len(var_chunks) - 1,
                 var_chunk_size=var_chunks[-1],
+                axis=axis,
                 name=name,
-                dim_i=dim_i,
                 chunk_size=chunk_size,
                 region=interval,
                 enc_chunks=enc_chunks,

xarray/core/dataset.py

@@ -2302,6 +2302,7 @@ def to_zarr(
             append_dim=append_dim,
             region=region,
             safe_chunks=safe_chunks,
+            align_chunks=align_chunks,
             zarr_version=zarr_version,
             zarr_format=zarr_format,
             write_empty_chunks=write_empty_chunks,

xarray/tests/test_backends.py

@@ -6868,6 +6868,54 @@ def test_zarr_safe_chunk_region(self, mode: Literal["r+", "a"]):
             chunk = chunk.chunk()
             self.save(store, chunk.chunk(), region=region)
 
+    @requires_dask
+    def test_dataset_to_zarr_align_chunks_true(self, tmp_store) -> None:
+        # This test is a replica of the one in `test_dataarray_to_zarr_align_chunks_true`
+        # but for datasets
+        with self.create_zarr_target() as store:
+            ds = (
+                DataArray(
+                    np.arange(4).reshape((2, 2)),
+                    dims=["a", "b"],
+                    coords={
+                        "a": np.arange(2),
+                        "b": np.arange(2),
+                    },
+                )
+                .chunk(a=(1, 1), b=(1, 1))
+                .to_dataset(name="foo")
+            )
+
+            self.save(
+                store,
+                ds,
+                align_chunks=True,
+                encoding={"foo": {"chunks": (3, 3)}},
+                mode="w",
+            )
+            assert_identical(ds, xr.open_zarr(store))
+
+            ds = (
+                DataArray(
+                    np.arange(4, 8).reshape((2, 2)),
+                    dims=["a", "b"],
+                    coords={
+                        "a": np.arange(2),
+                        "b": np.arange(2),
+                    },
+                )
+                .chunk(a=(1, 1), b=(1, 1))
+                .to_dataset(name="foo")
+            )
+
+            self.save(
+                store,
+                ds,
+                align_chunks=True,
+                region="auto",
+            )
+            assert_identical(ds, xr.open_zarr(store))
+
 
 @requires_h5netcdf
 @requires_fsspec