Support reading initial-defaults

Author: Fokko

pyiceberg/io/pyarrow.py

@@ -1768,9 +1768,12 @@ def struct(
                 array = self._cast_if_needed(field, field_array)
                 field_arrays.append(array)
                 fields.append(self._construct_field(field, array.type))
-            elif field.optional:
+            elif field.optional or field.initial_default is not None:
                 arrow_type = schema_to_pyarrow(field.field_type, include_field_ids=False)
-                field_arrays.append(pa.nulls(len(struct_array), type=arrow_type))
+                if field.initial_default is None:
+                    field_arrays.append(pa.nulls(len(struct_array), type=arrow_type))
+                else:
+                    field_arrays.append(pa.repeat(field.initial_default, len(struct_array)))
                 fields.append(self._construct_field(field, arrow_type))
             else:
                 raise ResolveError(f"Field is required, and could not be found in the file: {field}")

tests/io/test_pyarrow.py

@@ -2203,6 +2203,17 @@ def test_identity_partition_on_multi_columns() -> None:
         ) == arrow_table.sort_by([("born_year", "ascending"), ("n_legs", "ascending"), ("animal", "ascending")])
 
 
+def test_initial_value() -> None:
+    # Have some fake data, otherwise it will generate a table without records
+    data = pa.record_batch([pa.nulls(10, pa.int32())], ["some_field"])
+    result = _to_requested_schema(
+        Schema(NestedField(1, "so-true", BooleanType(), required=True, initial_default=True)), Schema(), data
+    )
+    assert result.column_names == ["so-true"]
+    for val in result[0]:
+        assert val.as_py() is True
+
+
 def test__to_requested_schema_timestamps(
     arrow_table_schema_with_all_timestamp_precisions: pa.Schema,
     arrow_table_with_all_timestamp_precisions: pa.Table,