Improve "variable not found" error message

xarray/core/dataset.py

@@ -1539,10 +1539,18 @@ def __getitem__(
 
         Indexing with a list of names will return a new ``Dataset`` object.
         """
+        from xarray.core.formatting import shorten_list_repr
+
         if utils.is_dict_like(key):
             return self.isel(**key)
         if utils.hashable(key):
-            return self._construct_dataarray(key)
+            try:
+                return self._construct_dataarray(key)
+            except KeyError as e:
+                raise KeyError(
+                    f"No variable named {key!r}. Variables on the dataset include {shorten_list_repr(list(self.variables.keys()), max_items=10)}"
+                ) from e
+
         if utils.iterable_of_hashable(key):
             return self._copy_listed(key)
         raise ValueError(f"Unsupported key-type {type(key)}")

xarray/core/formatting.py

@@ -4,9 +4,10 @@
 
 import contextlib
 import functools
+import itertools
 import math
 from collections import defaultdict
-from collections.abc import Collection, Hashable
+from collections.abc import Collection, Hashable, Sequence
 from datetime import datetime, timedelta
 from itertools import chain, zip_longest
 from reprlib import recursive_repr
@@ -937,3 +938,16 @@ def diff_dataset_repr(a, b, compat):
         summary.append(diff_attrs_repr(a.attrs, b.attrs, compat))
 
     return "\n".join(summary)
+
+
+def shorten_list_repr(items: Sequence, max_items: int) -> str:
+    if len(items) < max_items:
+        return repr(items)
+    else:
+        return repr(
+            list(
+                itertools.chain(
+                    items[: max_items // 2], ["..."], items[-max_items // 2 :]
+                )
+            )
+        )

xarray/tests/test_error_messages.py

@@ -0,0 +1,17 @@
+"""
+This new file is intended to test the quality & friendliness of error messages that are
+raised by xarray. It's currently separate from the standard tests, which are more
+focused on the functions working (though we could consider integrating them.).
+"""
+
+import pytest
+
+
+def test_no_var_in_dataset(ds):
+    with pytest.raises(
+        KeyError,
+        match=(
+            r"No variable named 'foo'. Variables on the dataset include \['z1', 'z2', 'x', 'time', 'c', 'y'\]"
+        ),
+    ):
+        ds["foo"]