Passing Empty List to multirange_index and df should return Empty…

… Results (#1412)

* `.df` and `.multi_index` Empty Sequence Returns Empty Results
* Use TypeVar for EmptyRange

---------

Co-authored-by: Agisilaos Kounelis <[email protected]>

tiledb/multirange_indexing.py

@@ -17,6 +17,7 @@
     Optional,
     Sequence,
     Tuple,
+    TypeVar,
     Union,
     cast,
 )
@@ -44,8 +45,17 @@
 current_timer: ContextVar[str] = ContextVar("timer_scope")
 
 
-# sentinel value to denote selecting an empty range
-EmptyRange = object()
+# sentinel type to denote selecting an empty range
+EmptyRange = TypeVar("EmptyRange")
+
+
+def is_empty_range(idx: Union[EmptyRange, List, Tuple]) -> bool:
+    if idx is not EmptyRange:
+        if hasattr(idx, "__len__") and len(idx) == 0 and idx != "":
+            return True
+        return False
+    return True
+
 
 # TODO: expand with more accepted scalar types
 Scalar = Real
@@ -248,7 +258,7 @@ def return_incomplete(self) -> bool:
 
     def __getitem__(self, idx):
         with timing("getitem_time"):
-            if idx is EmptyRange:
+            if is_empty_range(idx):
                 self.pyquery = None
                 self.subarray = None
             else:

tiledb/tests/test_multi_index-hp.py

@@ -39,7 +39,12 @@ def _direct_query_ranges(array: SparseArray, ranges, order):
         q.set_subarray(subarray)
 
         q.submit()
-    return {k: v[0].view(array.attr(0).dtype) for k, v in q.results().items()}
+
+    if ranges == [[]]:
+        # empty range should give empty result
+        return {k: [] for k in q.results()}
+    else:
+        return {k: v[0].view(array.attr(0).dtype) for k, v in q.results().items()}
 
 
 # Compound strategies to build valid inputs for multi_index
@@ -90,7 +95,6 @@ def create_array(uri):
     def test_multi_index_two_way_query(self, order, ranges, sparse_array_1d):
         """This test checks the result of "direct" range queries using PyQuery
         against the result of `multi_index` on the same ranges."""
-
         uri = sparse_array_1d
 
         assert isinstance(uri, str)

tiledb/tests/test_multi_index.py

@@ -903,6 +903,30 @@ def test_fixed_width_char(self):
         with tiledb.open(uri, mode="r") as A:
             assert all(A.query(use_arrow=True).df[:][""] == data)
 
+    @pytest.mark.skipif(not has_pandas(), reason="pandas>=1.0,<3.0 not installed")
+    def test_empty_idx(self):
+        uri = self.path("test_empty_idx")
+
+        schema = tiledb.ArraySchema(
+            domain=tiledb.Domain(tiledb.Dim(name="dim", domain=(0, 9), dtype=np.uint8)),
+            sparse=True,
+            attrs=[tiledb.Attr(name="a", dtype=np.float64)],
+        )
+        tiledb.Array.create(uri, schema)
+
+        data = np.array(np.random.randint(10, size=10), dtype=np.float64)
+
+        with tiledb.open(uri, mode="w") as A:
+            A[np.arange(10)] = data
+
+        with tiledb.open(uri, mode="r") as A:
+            assert_array_equal(A.df[tiledb.EmptyRange]["a"], [])
+            assert_array_equal(A.multi_index[tiledb.EmptyRange]["a"], [])
+            assert_array_equal(A.df[[]]["a"], [])
+            assert_array_equal(A.multi_index[[]]["a"], [])
+            assert_array_equal(A.df[()]["a"], [])
+            assert_array_equal(A.multi_index[()]["a"], [])
+
 
 # parametrize dtype and sparse
 @pytest.mark.parametrize(