Don't drop tables without data

dlt/pipeline/drop.py

@@ -114,7 +114,9 @@ def drop_resources(
         resource_pattern = None
 
     if resource_pattern:
-        data_tables = {t["name"]: t for t in schema.data_tables()}  # Don't remove _dlt tables
+        data_tables = {
+            t["name"]: t for t in schema.data_tables(seen_data_only=True)
+        }  # Don't remove _dlt tables
         resource_tables = group_tables_by_resource(data_tables, pattern=resource_pattern)
         resource_names = list(resource_tables.keys())
         # TODO: If drop_tables

tests/pipeline/test_refresh_modes.py

@@ -50,7 +50,11 @@ def some_data_3():
             yield {"id": 5, "name": "Jack"}
             yield {"id": 6, "name": "Jill"}
 
-        return [some_data_1, some_data_2, some_data_3]
+        @dlt.resource
+        def some_data_4():
+            yield []
+
+        return [some_data_1, some_data_2, some_data_3, some_data_4]
 
     # First run pipeline with load to destination so tables are created
     pipeline = dlt.pipeline(
@@ -70,6 +74,13 @@ def some_data_3():
     # pipeline.normalize()
     # pipeline.load()
 
+    assert set(t["name"] for t in pipeline.default_schema.data_tables(include_incomplete=True)) == {
+        "some_data_1",
+        "some_data_2",
+        # Table has never seen data and is not dropped
+        "some_data_4",
+    }
+
     # Confirm resource tables not selected on second run got wiped
     with pytest.raises(DatabaseUndefinedRelation):
         with pipeline.sql_client() as client: