fix: include schema temp table

sqlmesh/core/snapshot/evaluator.py

@@ -2185,7 +2185,13 @@ def _get_target_and_source_columns(
         if model.on_destructive_change.is_ignore or model.on_additive_change.is_ignore:
             # We need to identify the columns that are only in the source so we create an empty table with
             # the user query to determine that
-            with self.adapter.temp_table(model.ctas_query(**render_kwargs)) as temp_table:
+            temp_table_name = exp.table_(
+                "diff",
+                db=model.physical_schema,
+            )
+            with self.adapter.temp_table(
+                model.ctas_query(**render_kwargs), name=temp_table_name
+            ) as temp_table:
                 source_columns = list(self.adapter.columns(temp_table))
         else:
             source_columns = None

tests/core/test_snapshot_evaluator.py

@@ -3,6 +3,7 @@
 
 from typing_extensions import Self
 from unittest.mock import call, patch, Mock
+import contextlib
 import re
 import logging
 import pytest
@@ -2077,6 +2078,68 @@ def columns(table_name):
         )
 
 
+def test_temp_table_includes_schema_for_ignore_changes(
+    mocker: MockerFixture,
+    make_snapshot,
+    make_mocked_engine_adapter,
+):
+    """Test that temp table creation includes the physical schema when on_destructive_change or on_additive_change is IGNORE."""
+    # Create a model with on_destructive_change=IGNORE to trigger temp table creation
+    model = SqlModel(
+        name="test_schema.test_model",
+        kind=IncrementalByTimeRangeKind(
+            time_column="a", on_destructive_change=OnDestructiveChange.IGNORE
+        ),
+        query=parse_one("SELECT c, a FROM tbl WHERE ds BETWEEN @start_ds and @end_ds"),
+    )
+    snapshot = make_snapshot(model, version="1")
+    snapshot.categorize_as(SnapshotChangeCategory.BREAKING)
+
+    # Set up the mocked adapter
+    adapter = make_mocked_engine_adapter(EngineAdapter)
+    adapter.with_settings = lambda **kwargs: adapter  # type: ignore
+    adapter.table_exists = lambda _: True  # type: ignore
+
+    # Mock columns method to return existing columns
+    def columns(table_name):
+        return {
+            "c": exp.DataType.build("int"),
+            "a": exp.DataType.build("int"),
+        }
+
+    adapter.columns = columns  # type: ignore
+
+    # Create a mock for the temp_table context manager
+    temp_table_name_captured = None
+
+    @contextlib.contextmanager
+    def mock_temp_table(query_or_df, name="diff", **kwargs):
+        nonlocal temp_table_name_captured
+        temp_table_name_captured = exp.to_table(name)
+        # Return a table that temp_table would normally return
+        yield exp.table_("__temp_diff_12345", db=temp_table_name_captured.db)
+
+    adapter.temp_table = mock_temp_table  # type: ignore
+    adapter.insert_append = lambda *args, **kwargs: None  # type: ignore
+
+    evaluator = SnapshotEvaluator(adapter)
+
+    # Call the append method which will trigger _get_target_and_source_columns
+    evaluator.evaluate(
+        snapshot,
+        start="2020-01-01",
+        end="2020-01-02",
+        execution_time="2020-01-02",
+        snapshots={},
+    )
+
+    # Verify that temp_table was called with a name that includes the schema
+    assert temp_table_name_captured is not None
+    assert temp_table_name_captured.name == "diff"
+    assert temp_table_name_captured.db == model.physical_schema
+    assert str(temp_table_name_captured.db) == "sqlmesh__test_schema"
+
+
 def test_forward_only_snapshot_for_added_model(mocker: MockerFixture, adapter_mock, make_snapshot):
     adapter_mock.SUPPORTS_CLONING = False
     evaluator = SnapshotEvaluator(adapter_mock)