fix: Validate SetStatisticsUpdate correctly (fixes #2865)

Previously the pydantic @model_validator would fail because it assumed
statistics was a model instance. In a "before"" validator that is not
necessarily the case.

Check type explicitly with isinstance instead, and handle `dict` case
too.

Author: ragnard

pyiceberg/table/update/__init__.py

@@ -21,7 +21,7 @@
 from abc import ABC, abstractmethod
 from datetime import datetime
 from functools import singledispatch
-from typing import TYPE_CHECKING, Annotated, Any, Generic, Literal, TypeVar, cast
+from typing import TYPE_CHECKING, Annotated, Any, Generic, Literal, TypeVar
 
 from pydantic import Field, field_validator, model_serializer, model_validator
 
@@ -181,9 +181,15 @@ class SetStatisticsUpdate(IcebergBaseModel):
 
     @model_validator(mode="before")
     def validate_snapshot_id(cls, data: dict[str, Any]) -> dict[str, Any]:
-        stats = cast(StatisticsFile, data["statistics"])
+        snapshot_id = None
 
-        data["snapshot_id"] = stats.snapshot_id
+        stats = data["statistics"]
+        if isinstance(stats, StatisticsFile):
+            snapshot_id = stats.snapshot_id
+        elif isinstance(stats, dict):
+            snapshot_id = stats.get("snapshot_id")
+
+        data["snapshot_id"] = snapshot_id
 
         return data
 

tests/table/test_init.py

@@ -21,7 +21,7 @@
 from typing import Any
 
 import pytest
-from pydantic import ValidationError
+from pydantic import ValidationError, BaseModel
 from sortedcontainers import SortedList
 
 from pyiceberg.catalog.noop import NoopCatalog
@@ -1391,6 +1391,8 @@ def test_set_statistics_update(table_v2_with_statistics: Table) -> None:
         statistics=statistics_file,
     )
 
+    assert model_roundtrips(update)
+
     new_metadata = update_table_metadata(
         table_v2_with_statistics.metadata,
         (update,),
@@ -1575,3 +1577,14 @@ def test_add_snapshot_update_updates_next_row_id(table_v3: Table) -> None:
 
     new_metadata = update_table_metadata(table_v3.metadata, (AddSnapshotUpdate(snapshot=new_snapshot),))
     assert new_metadata.next_row_id == 11
+
+
+def model_roundtrips(model: BaseModel) -> bool:
+    """Helper assertion that tests if a pydantic model roundtrips
+    successfully.
+    """
+    __tracebackhide__ = True
+    model_data = model.model_dump()
+    if model != type(model).model_validate(model_data):
+        pytest.fail(f"model {type(model)} did not roundtrip successfully")
+    return True