Reject empty source-ids in PartitionField / SortField

pyiceberg/partitioning.py

@@ -109,7 +109,8 @@ def __init__(
     @classmethod
     def map_source_ids_onto_source_id(cls, data: Any) -> Any:
         if isinstance(data, dict):
-            if "source-id" not in data and (source_ids := data["source-ids"]):
+            if "source-id" not in data and "source-ids" in data:
+                source_ids = data["source-ids"]
                 if isinstance(source_ids, list):
                     if len(source_ids) == 0:
                         raise ValueError("Empty source-ids is not allowed")

pyiceberg/table/sorting.py

@@ -101,7 +101,8 @@ def set_null_order(cls, values: dict[str, Any]) -> dict[str, Any]:
     @classmethod
     def map_source_ids_onto_source_id(cls, data: Any) -> Any:
         if isinstance(data, dict):
-            if "source-id" not in data and (source_ids := data["source-ids"]):
+            if "source-id" not in data and "source-ids" in data:
+                source_ids = data["source-ids"]
                 if isinstance(source_ids, list):
                     if len(source_ids) == 0:
                         raise ValueError("Empty source-ids is not allowed")

tests/table/test_partitioning.py

@@ -267,6 +267,12 @@ def test_deserialize_partition_field_v3() -> None:
     assert field == PartitionField(source_id=1, field_id=1000, transform=TruncateTransform(width=19), name="str_truncate")
 
 
+def test_deserialize_partition_field_empty_source_ids_rejected() -> None:
+    json_partition_spec = """{"source-ids": [], "field-id": 1000, "transform": "identity", "name": "x"}"""
+    with pytest.raises(Exception, match="Empty source-ids is not allowed"):
+        PartitionField.model_validate_json(json_partition_spec)
+
+
 def test_incompatible_source_column_not_found() -> None:
     schema = Schema(NestedField(1, "foo", IntegerType()), NestedField(2, "bar", IntegerType()))
 

tests/table/test_sorting.py

@@ -138,6 +138,12 @@ def test_serialize_sort_field_v3() -> None:
     assert SortField.model_validate_json(payload) == expected
 
 
+def test_deserialize_sort_field_empty_source_ids_rejected() -> None:
+    payload = '{"source-ids":[],"transform":"identity","direction":"asc","null-order":"nulls-first"}'
+    with pytest.raises(Exception, match="Empty source-ids is not allowed"):
+        SortField.model_validate_json(payload)
+
+
 def test_incompatible_source_column_not_found(sort_order: SortOrder) -> None:
     schema = Schema(NestedField(1, "foo", IntegerType()), NestedField(2, "bar", IntegerType()))