Fix for zarr encoding bug (#2320)

* simple fix for zarr encoding bug

* remove test asserting error raised for bad chunk sizes

* bump for ci

Author: Joe Hamman

doc/whats-new.rst

@@ -61,6 +61,10 @@ Bug fixes
   attribute being set.
   (:issue:`2201`)
   By `Thomas Voigt <https://github.com/tv3141>`_.
+- Fixed a bug in ``zarr`` backend which prevented use with datasets with
+  invalid chunk size encoding after reading from an existing store
+  (:issue:`2278`).
+  By `Joe Hamman <https://github.com/jhamman>`_.
 
 - Tests can be run in parallel with pytest-xdist
   By `Tony Tung <https://github.com/ttung>`_.

xarray/backends/zarr.py

@@ -102,9 +102,8 @@ def _determine_zarr_chunks(enc_chunks, var_chunks, ndim):
         enc_chunks_tuple = tuple(enc_chunks)
 
     if len(enc_chunks_tuple) != ndim:
-        raise ValueError("zarr chunks tuple %r must have same length as "
-                         "variable.ndim %g" %
-                         (enc_chunks_tuple, ndim))
+        # throw away encoding chunks, start over
+        return _determine_zarr_chunks(None, var_chunks, ndim)
 
     for x in enc_chunks_tuple:
         if not isinstance(x, int):

xarray/tests/test_backends.py

@@ -1404,12 +1404,6 @@ def test_chunk_encoding_with_dask(self):
             with self.roundtrip(ds_chunk4) as actual:
                 self.assertEqual((4,), actual['var1'].encoding['chunks'])
 
-        # specify incompatible encoding
-        ds_chunk4['var1'].encoding.update({'chunks': (5, 5)})
-        with pytest.raises(ValueError) as e_info:
-            with self.roundtrip(ds_chunk4) as actual:
-                pass
-        assert e_info.match('chunks')
 
         # TODO: remove this failure once syncronized overlapping writes are
         # supported by xarray
@@ -1522,6 +1516,21 @@ def test_to_zarr_compute_false_roundtrip(self):
             with self.open(store) as actual:
                 assert_identical(original, actual)
 
+    def test_encoding_chunksizes(self):
+        # regression test for GH2278
+        # see also test_encoding_chunksizes_unlimited
+        nx, ny, nt = 4, 4, 5
+        original = xr.Dataset({}, coords={'x': np.arange(nx),
+                                          'y': np.arange(ny),
+                                          't': np.arange(nt)})
+        original['v'] = xr.Variable(('x', 'y', 't'), np.zeros((nx, ny, nt)))
+        original = original.chunk({'t': 1, 'x': 2, 'y': 2})
+
+        with self.roundtrip(original) as ds1:
+            assert_equal(ds1, original)
+            with self.roundtrip(ds1.isel(t=0)) as ds2:
+                assert_equal(ds2, original.isel(t=0))
+
 
 @requires_zarr
 class ZarrDictStoreTest(BaseZarrTest, TestCase):