This is an automated email from the ASF dual-hosted git repository. sungwy pushed a commit to branch fix-null-identity-partition-default in repository https://gitbox.apache.org/repos/asf/iceberg-python.git
commit 491042af8c82f6b908405d8b1439af3784686ae5 Author: Sung Yun (CODE SIGNING KEY) <[email protected]> AuthorDate: Tue Oct 6 17:30:24 2026 -0400 fix column projection edge case with null identity partition --- pyiceberg/io/pyarrow.py | 8 +++---- tests/io/test_pyarrow.py | 57 ++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 60 insertions(+), 5 deletions(-) diff --git a/pyiceberg/io/pyarrow.py b/pyiceberg/io/pyarrow.py index 2dcb8a579..df37b8dd1 100644 --- a/pyiceberg/io/pyarrow.py +++ b/pyiceberg/io/pyarrow.py @@ -1641,9 +1641,7 @@ def _get_column_projection_values( for field_id in project_schema_diff: for partition_field in partition_spec.fields_by_source_id(field_id): if isinstance(partition_field.transform, IdentityTransform): - partition_value = accessors[partition_field.field_id].get(file.partition) - if partition_value is not None: - projected_missing_fields[field_id] = partition_value + projected_missing_fields[field_id] = accessors[partition_field.field_id].get(file.partition) return projected_missing_fields @@ -2071,8 +2069,8 @@ class ArrowProjectionVisitor(SchemaWithPartnerVisitor[pa.Array, pa.Array | None] elif field.optional or field.initial_default is not None: # When an optional field is added, or when a required field with a non-null initial default is added arrow_type = schema_to_pyarrow(field.field_type, include_field_ids=self._include_field_ids) - projected_value = self._projected_missing_fields.get(field.field_id) - if projected_value is not None: + if field.field_id in self._projected_missing_fields: + projected_value = self._projected_missing_fields[field.field_id] field_arrays.append(pa.repeat(pa.scalar(projected_value, type=arrow_type), len(struct_array))) elif field.initial_default is None: field_arrays.append(pa.nulls(len(struct_array), type=arrow_type)) diff --git a/tests/io/test_pyarrow.py b/tests/io/test_pyarrow.py index 4d5d4431c..147f8b45d 100644 --- a/tests/io/test_pyarrow.py +++ b/tests/io/test_pyarrow.py @@ -58,7 +58,9 @@ from pyiceberg.expressions import ( BoundNotStartsWith, BoundReference, BoundStartsWith, + EqualTo, GreaterThan, + IsNull, Not, Or, ) @@ -1460,6 +1462,61 @@ def test_identity_transform_column_projection_with_falsy_value( ) [email protected]( + "row_filter, expected_other_field", + [ + (None, ["foo", "bar"]), + (IsNull("partition_col"), ["foo", "bar"]), + (EqualTo("partition_col", 7), []), + ], +) +def test_identity_transform_null_partition_value_takes_precedence_over_initial_default( + tmpdir: str, row_filter: BooleanExpression | None, expected_other_field: list[str] +) -> None: + """A null identity partition value must project as null, not as the field's initial-default.""" + from pyiceberg.expressions.visitors import bind + + table_schema = Schema( + NestedField(1, "other_field", StringType(), required=False), + NestedField(2, "partition_col", IntegerType(), required=False, initial_default=7), + ) + partition_spec = PartitionSpec(PartitionField(2, 1000, IdentityTransform(), "partition_col")) + + arrow_schema = pa.schema([pa.field("other_field", pa.string(), metadata={PYARROW_PARQUET_FIELD_ID_KEY: "1"})]) + file_path = _write_table_to_file( + f"{tmpdir}/null_partition.parquet", arrow_schema, pa.table({"other_field": ["foo", "bar"]}, schema=arrow_schema) + ) + data_file = DataFile.from_args( + content=DataFileContent.DATA, + file_path=file_path, + file_format=FileFormat.PARQUET, + partition=Record(None), + record_count=2, + file_size_in_bytes=os.path.getsize(file_path), + spec_id=partition_spec.spec_id, + ) + + batches = list( + _task_to_record_batches( + PyArrowFileIO(), + FileScanTask(data_file), + bound_row_filter=AlwaysTrue() if row_filter is None else bind(table_schema, row_filter, case_sensitive=True), + projected_schema=table_schema, + table_schema=table_schema, + projected_field_ids={1, 2}, + positional_deletes=None, + case_sensitive=True, + partition_spec=partition_spec, + format_version=3, + ) + ) + + assert [value for batch in batches for value in batch.column("other_field").to_pylist()] == expected_other_field + assert [value for batch in batches for value in batch.column("partition_col").to_pylist()] == [None] * len( + expected_other_field + ) + + def test_identity_transform_columns_projection(tmp_path: str, catalog: InMemoryCatalog) -> None: # Test by adding a non-partitioned data file to a multi-partitioned table, verifying partition value # projection from manifest metadata.
