Update contains_cftime_datetimes to avoid loading entire variable array

doc/whats-new.rst

@@ -55,6 +55,8 @@ Bug fixes
   By `Kai Mühlbauer <https://github.com/kmuehlbauer>`_ and `Scott Chamberlin <https://github.com/scottcha>`_.
 - Handle ``keep_attrs`` option in binary operators of :py:meth:`Dataset` (:issue:`7390`, :pull:`7391`).
   By `Aron Gergely <https://github.com/arongergely>`_.
+- Improved performance in ``open_dataset`` for datasets with large object arrays (:issue:`7484`, :pull:`7494`).
+  By `Alex Goodman <https://github.com/agoodm>`_.
 
 Documentation
 ~~~~~~~~~~~~~

xarray/core/accessor_dt.py

@@ -575,7 +575,7 @@ def __new__(cls, obj: T_DataArray) -> CombinedDatetimelikeAccessor:
         # we need to choose which parent (datetime or timedelta) is
         # appropriate. Since we're checking the dtypes anyway, we'll just
         # do all the validation here.
-        if not _contains_datetime_like_objects(obj):
+        if not _contains_datetime_like_objects(obj.variable):
             raise TypeError(
                 "'.dt' accessor only available for "
                 "DataArray with datetime64 timedelta64 dtype or "

xarray/core/common.py

@@ -40,6 +40,7 @@
         ScalarOrArray,
         SideOptions,
         T_DataWithCoords,
+        T_Variable,
     )
     from xarray.core.variable import Variable
 
@@ -1771,31 +1772,37 @@ def is_np_timedelta_like(dtype: DTypeLike) -> bool:
     return np.issubdtype(dtype, np.timedelta64)
 
 
-def _contains_cftime_datetimes(array) -> bool:
+def _contains_cftime_datetimes(array: Any) -> bool:
     """Check if an array contains cftime.datetime objects"""
-    if cftime is None:
-        return False
+    from xarray.core.variable import Variable
+
+    if isinstance(array, Variable):
+        var = array
     else:
-        if array.dtype == np.dtype("O") and array.size > 0:
-            sample = np.asarray(array).flat[0]
-            if is_duck_dask_array(sample):
-                sample = sample.compute()
-                if isinstance(sample, np.ndarray):
-                    sample = sample.item()
-            return isinstance(sample, cftime.datetime)
-        else:
-            return False
+        var = Variable(dims=tuple(f"dim_{v}" for v in range(array.ndim)), data=array)
+
+    return _variable_contains_cftime_datetimes(var)
 
 
-def contains_cftime_datetimes(var) -> bool:
+def _variable_contains_cftime_datetimes(var: T_Variable) -> bool:
     """Check if an xarray.Variable contains cftime.datetime objects"""
-    if var.dtype == np.dtype("O") and var.size > 0:
-        return _contains_cftime_datetimes(var.data)
-    else:
+    if cftime is None:
         return False
 
+    if var.dtype == np.dtype("O") and var.size > 0:
+        first_idx = tuple(0 for _ in range(var.ndim))
+        sample = var[first_idx]
+        return isinstance(sample.to_numpy().item(), cftime.datetime)
+
+    return False
+
+
+def contains_cftime_datetimes(var: T_Variable) -> bool:
+    """Check if an xarray.Variable contains cftime.datetime objects"""
+    return _variable_contains_cftime_datetimes(var)
+
 
-def _contains_datetime_like_objects(var) -> bool:
+def _contains_datetime_like_objects(var: T_Variable) -> bool:
     """Check if a variable contains datetime like objects (either
     np.datetime64, np.timedelta64, or cftime.datetime)
     """