fix: add workaround for NULL map bugs

partially addresses #8632

ibis/backends/tests/test_map.py

@@ -26,6 +26,22 @@
 ]
 
 
+@pytest.mark.parametrize(
+    "kv",
+    [
+        param(([1, 2], None), id="null_values"),
+        param((None, [1, 2]), id="null_keys"),
+        param((None, None), id="null_both"),
+    ],
+)
+def test_map_nulls(con, kv):
+    k, v = kv
+    k = ibis.literal(k, type="array<int64>")
+    v = ibis.literal(v, type="array<int64>")
+    m = ibis.map(k, v)
+    assert con.execute(m) is None
+
+
 @pytest.mark.notimpl(["pandas", "dask"])
 def test_map_table(backend):
     table = backend.map

ibis/expr/types/maps.py

@@ -4,9 +4,11 @@
 
 from public import public
 
+import ibis
 import ibis.expr.operations as ops
 import ibis.expr.types as ir
 from ibis.common.deferred import deferrable
+from ibis.expr import datatypes as dt
 from ibis.expr.types.generic import Column, Scalar, Value
 
 if TYPE_CHECKING:
@@ -488,4 +490,17 @@ def map(
         return keys
     if values is None:
         keys, values = tuple(keys.keys()), tuple(keys.values())
-    return ops.Map(keys, values).to_expr()
+
+    # workaround for https://github.com/ibis-project/ibis/issues/8632
+    key_array = ibis.array(keys)
+    value_array = ibis.array(values)
+    regular = ops.Map(key_array, value_array).to_expr()
+    either_null = key_array.isnull() | value_array.isnull()
+    null = ibis.literal(
+        None,
+        type=dt.Map(
+            key_type=key_array.type().value_type,
+            value_type=value_array.type().value_type,
+        ),
+    )
+    return either_null.ifelse(null, regular)