⚡️ (store/sql) Don't use inner query if not needed

.pre-commit-config.yaml

@@ -46,7 +46,7 @@ repos:
     hooks:
       - id: pyproject-flake8
         additional_dependencies: [ flake8-bugbear ]
-        args: [ "--extend-ignore", "E203, E501" ]
+        args: [ "--extend-ignore", "E203, E501, W503" ]
         exclude: (test_[\w]+\.py|\.csv|\.json|\.lock)$
 
   - repo: https://github.com/codespell-project/codespell

ftmq/sql.py

@@ -147,11 +147,15 @@ def all_canonical_ids(self) -> Select:
 
     @cached_property
     def _unsorted_statements(self) -> Select:
-        return (
-            select(self.table)
-            .where(self.table.c.canonical_id.in_(self.canonical_ids))
-            .order_by(self.table.c.canonical_id)
-        )
+        where = self.clause
+        if (
+            self.q.properties
+            or self.q.reversed
+            or self.q.search_filters
+            or self.q.limit
+        ):
+            where = self.table.c.canonical_id.in_(self.canonical_ids)
+        return select(self.table).where(where).order_by(self.table.c.canonical_id)
 
     @cached_property
     def _sorted_statements(self) -> Select:

setup.cfg

@@ -1,4 +1,4 @@
 [flake8]
 max-line-length = 88
 select = C,E,F,W,B,B950
-extend-ignore = E203, E501
+extend-ignore = E203, E501, W503

tests/test_sql.py

@@ -335,6 +335,35 @@ def test_sql():
         """,
     )
 
+    # simplified if no properties, reversed or search:
+    q = Query()
+    assert _compare_str(
+        q.sql.statements,
+        """
+        SELECT test_table.id, test_table.entity_id, test_table.canonical_id,
+        test_table.prop, test_table.prop_type, test_table.schema, test_table.value,
+        test_table.original_value, test_table.dataset, test_table.lang,
+        test_table.target, test_table.external, test_table.first_seen,
+        test_table.last_seen FROM test_table ORDER BY test_table.canonical_id
+        """,
+    )
+    q = Query().where(dataset="foo", schema="Person")
+    assert _compare_str(
+        q.sql.statements,
+        """
+        SELECT test_table.id, test_table.entity_id, test_table.canonical_id,
+        test_table.prop, test_table.prop_type, test_table.schema,
+        test_table.value, test_table.original_value, test_table.dataset,
+        test_table.lang, test_table.target, test_table.external,
+        test_table.first_seen, test_table.last_seen FROM test_table WHERE
+        test_table.dataset = :dataset_1 AND test_table.schema = :schema_1 ORDER
+        BY test_table.canonical_id
+        """,
+    )
+
+    # but we need complex query if we want a limit:
+    assert "canonical_id IN" in str(q[:10].sql.statements)
+
 
 def test_sql_search_query():
     q = Query().search("agency", ["name"])

tests/test_store.py

@@ -56,6 +56,12 @@ def _run_store_test(cls: Store, proxies, **kwargs):
         break
     assert tested
 
+    # iterate
+    entities = [e for e in store.iterate()]
+    assert len(entities) == 474 + 151
+    entities = [e for e in store.iterate(dataset="eu_authorities")]
+    assert len(entities) == 151
+
     view = store.default_view()
     ds = make_dataset("eu_authorities")
     view = store.view(ds)