add test and ignore type checks (#82)

Co-authored-by: aandres3 <[email protected]>

CHANGELOG.md

@@ -6,6 +6,14 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/)
 and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
 
 <!-- insertion marker -->
+## [v0.8.0](https://github.com/tradewelltech/protarrow/releases/tag/v0.8.0) - 2024-10-30
+
+<small>[Compare with v0.7.0](https://github.com/tradewelltech/protarrow/compare/v0.7.0...v0.8.0)</small>
+
+### Added
+
+- Add ignore for type check ([08f591f](https://github.com/tradewelltech/protarrow/commit/08f591f5625f0cf3bd43ba2f61cb3c0a1795c0e2) by aandres3).
+
 ## [v0.7.0](https://github.com/tradewelltech/protarrow/releases/tag/v0.7.0) - 2024-10-14
 
 <small>[Compare with v0.6.0](https://github.com/tradewelltech/protarrow/compare/v0.6.0...v0.7.0)</small>

protarrow/proto_to_arrow.py

@@ -520,15 +520,17 @@ def _messages_to_array(
                 nullable=_proto_field_nullable(field_descriptor, config),
             )
         )
-
+    if validity_mask is not None:
+        mask = pc.invert(pa.array(validity_mask, pa.bool_()))
+    elif len(arrays) == 0:
+        # This only happens when using empty messages.
+        mask = pa.repeat(False, len(messages))  # type: ignore[arg-type]
+    else:
+        mask = None
     return pa.StructArray.from_arrays(
         arrays=arrays,
         fields=fields,
-        mask=(
-            pc.invert(pa.array(validity_mask, pa.bool_()))
-            if validity_mask is not None
-            else pa.repeat(False, len(messages))
-        ),
+        mask=mask,
     )
 
 

tests/test_conversion.py

@@ -28,6 +28,7 @@
 from protarrow.message_extractor import MessageExtractor
 from protarrow.proto_to_arrow import (
     NestedIterable,
+    _messages_to_array,
     _repeated_proto_to_array,
     field_descriptor_to_field,
     message_type_to_schema,
@@ -199,6 +200,27 @@ def test_enum_values_as_int():
     assert array.to_pylist() == [[0, 1, 0], [], []]
 
 
+def test_messages_to_array_empty():
+    assert _messages_to_array(
+        [Empty(), Empty()], Empty.DESCRIPTOR, None, ProtarrowConfig()
+    ) == pa.StructArray.from_arrays([], names=[], mask=pa.array([False, False]))
+
+
+def test_empty_values():
+    records = [
+        ExampleMessage(empty_values=[Empty(), Empty(), Empty()]),
+        ExampleMessage(empty_values=[]),
+        ExampleMessage(),
+    ]
+
+    array = _repeated_proto_to_array(
+        NestedIterable(records, lambda x: x.empty_values),
+        ExampleMessage.DESCRIPTOR.fields_by_name["empty_values"],
+        ProtarrowConfig(),
+    )
+    assert array.to_pylist() == [[{}, {}, {}], [], []]
+
+
 @pytest.mark.parametrize(
     ["config", "expected"],
     [