ARROW-241 Allow list in Schema as an alias for pa.list_ (#223)

Author: lazargugleta

bindings/python/pymongoarrow/types.py

@@ -288,7 +288,10 @@ def _normalize_typeid(typeid, field_name):
             fields.append((sub_field_name, _normalize_typeid(sub_typeid, sub_field_name)))
         return struct(fields)
     if isinstance(typeid, list):
-        return list_(_normalize_typeid(type(typeid[0]), "0"))
+        if len(typeid) != 1:
+            msg = f"list field in schema must contain exactly one element, not {len(typeid)}"
+            raise ValueError(msg)
+        return list_(_normalize_typeid(typeid[0], "0"))
     if _is_typeid_supported(typeid):
         normalizer = _TYPE_NORMALIZER_FACTORY[typeid]
         return normalizer(typeid)

bindings/python/test/test_schema.py

@@ -14,6 +14,7 @@
 from datetime import datetime
 from unittest import TestCase
 
+import pytest
 from bson import Binary, Code, Decimal128, Int64, ObjectId
 from pyarrow import Table, field, float64, int64, list_, struct, timestamp
 from pyarrow import schema as ArrowSchema
@@ -94,3 +95,22 @@ def test_list_of_list_projection(self):
             }
         )
         self.assertEqual(schema._get_projection(), {"_id": True, "list": {"a": True, "b": True}})
+
+    def test_py_list_projection(self):
+        schema = Schema(
+            {"_id": ObjectId, "list": [(struct([field("a", int64()), field("b", float64())]))]}
+        )
+
+        self.assertEqual(schema._get_projection(), {"_id": True, "list": {"a": True, "b": True}})
+
+    def test_py_list_with_multiple_fields_raises(self):
+        with pytest.raises(
+            ValueError, match="list field in schema must contain exactly one element, not 2"
+        ):
+            _ = Schema({"_id": ObjectId, "list": [([field("a", int64()), field("b", float64())])]})
+
+    def test_py_empty_list_raises(self):
+        with pytest.raises(
+            ValueError, match="list field in schema must contain exactly one element, not 0"
+        ):
+            _ = Schema({"_id": ObjectId, "list": []})