fix schema_id during AddSchemaUpdate

Author: kevinjqliu

pyiceberg/table/__init__.py

@@ -504,8 +504,11 @@ def _(update: AddSchemaUpdate, base_metadata: TableMetadata, context: _TableMeta
     if update.last_column_id < base_metadata.last_column_id:
         raise ValueError(f"Invalid last column id {update.last_column_id}, must be >= {base_metadata.last_column_id}")
 
-    # PROBLEM: `update.schema_`'s `schema_id` starts with 0 but should be 1
-    # fresh_schema = assign_fresh_schema_ids(update.schema_)
+    # `update.schema_.schema_id` should be the last_schema_id + 1
+    last_schema_id = max(schema.schema_id for schema in base_metadata.schemas)
+    next_schema_id = last_schema_id + 1
+    new_schema = update.schema_.model_copy(update={"schema_id": next_schema_id})
+    update = update.model_copy(update={"schema_": new_schema})
 
     context.add_update(update)
     return base_metadata.model_copy(

tests/catalog/test_base.py

@@ -397,8 +397,10 @@ def test_commit_table(catalog: InMemoryCatalog) -> None:
 
     # Then
     assert response.metadata.table_uuid == given_table.metadata.table_uuid
-    # assert len(response.metadata.schemas) == 1
-    # assert response.metadata.schemas[0] == new_schema
+    assert given_table.metadata.current_schema_id == 1
+    assert len(response.metadata.schemas) == 2
+    assert response.metadata.schemas[1] == new_schema
+    assert given_table.metadata.last_column_id == new_schema.highest_field_id
 
 
 def test_add_column(catalog: InMemoryCatalog) -> None: