make validator reusable

Author: kevinjqliu

pyiceberg/catalog/rest.py

@@ -28,7 +28,7 @@
     Union,
 )
 
-from pydantic import Field, ValidationError
+from pydantic import Field, ValidationError, field_validator
 from requests import HTTPError, Session
 from tenacity import RetryCallState, retry, retry_if_exception_type, stop_after_attempt
 
@@ -69,6 +69,7 @@
 )
 from pyiceberg.table.sorting import UNSORTED_SORT_ORDER, SortOrder, assign_fresh_sort_order_ids
 from pyiceberg.typedef import EMPTY_DICT, UTF8, IcebergBaseModel
+from pyiceberg.types import transform_dict_value_to_str
 
 if TYPE_CHECKING:
     import pyarrow as pa
@@ -146,6 +147,8 @@ class CreateTableRequest(IcebergBaseModel):
     write_order: Optional[SortOrder] = Field(alias="write-order")
     stage_create: bool = Field(alias="stage-create", default=False)
     properties: Properties = Field(default_factory=dict)
+    # validators
+    transform_properties_dict_value_to_str = field_validator('properties', mode='before')(transform_dict_value_to_str)
 
 
 class RegisterTableRequest(IcebergBaseModel):

pyiceberg/table/metadata.py

@@ -49,6 +49,7 @@
     IcebergRootModel,
     Properties,
 )
+from pyiceberg.types import transform_dict_value_to_str
 from pyiceberg.utils.datetime import datetime_to_millis
 
 CURRENT_SNAPSHOT_ID = "current-snapshot-id"
@@ -178,15 +179,6 @@ class TableMetadataCommonFields(IcebergBaseModel):
     to be used for arbitrary metadata. For example, commit.retry.num-retries
     is used to control the number of commit retries."""
 
-    @field_validator("properties", mode='before')
-    @classmethod
-    def transform_dict_value_to_str(cls, dict: Dict[str, Any]) -> Dict[str, str]:
-        """Transform all values in the dictionary to string. Raise an error if any value is None."""
-        for value in dict.values():
-            if value is None:
-                raise ValueError("None type is not a supported value in properties")
-        return {k: str(v) for k, v in dict.items()}
-
     current_snapshot_id: Optional[int] = Field(alias="current-snapshot-id", default=None)
     """ID of the current table snapshot."""
 
@@ -235,6 +227,9 @@ def schema_by_id(self, schema_id: int) -> Optional[Schema]:
         """Get the schema by schema_id."""
         return next((schema for schema in self.schemas if schema.schema_id == schema_id), None)
 
+    # validators
+    transform_properties_dict_value_to_str = field_validator('properties', mode='before')(transform_dict_value_to_str)
+
 
 class TableMetadataV1(TableMetadataCommonFields, IcebergBaseModel):
     """Represents version 1 of the Table Metadata.

pyiceberg/types.py

@@ -37,6 +37,7 @@
 from typing import (
     Any,
     ClassVar,
+    Dict,
     Literal,
     Optional,
     Tuple,
@@ -61,6 +62,14 @@
 FIXED_PARSER = ParseNumberFromBrackets(FIXED)
 
 
+def transform_dict_value_to_str(dict: Dict[str, Any]) -> Dict[str, str]:
+    """Transform all values in the dictionary to string. Raise an error if any value is None."""
+    for value in dict.values():
+        if value is None:
+            raise ValueError("None type is not a supported value in properties")
+    return {k: str(v) for k, v in dict.items()}
+
+
 def _parse_decimal_type(decimal: Any) -> Tuple[int, int]:
     if isinstance(decimal, str):
         matches = DECIMAL_REGEX.search(decimal)

tests/integration/test_writes.py

@@ -28,12 +28,13 @@
 import pytest
 import pytz
 from pyarrow.fs import S3FileSystem
+from pydantic_core import ValidationError
 from pyspark.sql import SparkSession
 from pytest_mock.plugin import MockerFixture
 
 from pyiceberg.catalog import Catalog, Properties, Table, load_catalog
 from pyiceberg.catalog.sql import SqlCatalog
-from pyiceberg.exceptions import NamespaceAlreadyExistsError, NoSuchTableError, ServerError
+from pyiceberg.exceptions import NamespaceAlreadyExistsError, NoSuchTableError
 from pyiceberg.schema import Schema
 from pyiceberg.table import _dataframe_to_data_files
 from pyiceberg.types import (
@@ -705,8 +706,8 @@ def test_table_properties_raise_for_none_value(
     property_with_none = {"property_name": None}
     identifier = "default.test_table_properties_raise_for_none_value"
 
-    with pytest.raises(ServerError) as exc_info:
+    with pytest.raises(ValidationError) as exc_info:
         _ = _create_table(
             session_catalog, identifier, {"format-version": format_version, **property_with_none}, [arrow_table_with_null]
         )
-    assert "NullPointerException: null value in entry: property_name=null" in str(exc_info.value)
+    assert "None type is not a supported value in properties" in str(exc_info.value)