Fixes #6767: correctly rehash queries in a migration

migrations/versions/9e8c841d1a30_fix_hash.py

@@ -0,0 +1,64 @@
+"""fix_hash
+
+Revision ID: 9e8c841d1a30
+Revises: 7205816877ec
+Create Date: 2024-10-05 18:55:35.730573
+
+"""
+import logging
+from alembic import op
+import sqlalchemy as sa
+from sqlalchemy.sql import table
+from sqlalchemy import select
+
+from redash.query_runner import BaseQueryRunner, get_query_runner
+
+
+# revision identifiers, used by Alembic.
+revision = '9e8c841d1a30'
+down_revision = '7205816877ec'
+branch_labels = None
+depends_on = None
+
+
+def update_query_hash(record):
+    should_apply_auto_limit = record['options'].get("apply_auto_limit", False) if record['options'] else False
+    query_runner = get_query_runner(record['type'], {}) if record['type'] else BaseQueryRunner({})
+    query_text = record['query']
+
+    parameters_dict = {p["name"]: p.get("value") for p in record['options'].get('parameters', [])} if record.options else {}
+    if any(parameters_dict):
+        print(f"Query {record['query_id']} has parameters. Hash might be incorrect.")
+
+    return query_runner.gen_query_hash(query_text, should_apply_auto_limit)
+
+
+def upgrade():
+    conn = op.get_bind()
+
+    metadata = sa.MetaData(bind=conn)
+    queries = sa.Table("queries", metadata, autoload=True)
+    data_sources = sa.Table("data_sources", metadata, autoload=True)
+
+    joined_table = queries.outerjoin(data_sources, queries.c.data_source_id == data_sources.c.id)
+
+    query = select([
+        queries.c.id.label("query_id"),
+        queries.c.query,
+        queries.c.query_hash,
+        queries.c.options,
+        data_sources.c.id.label("data_source_id"),
+        data_sources.c.type
+    ]).select_from(joined_table)
+
+    for record in conn.execute(query):
+        new_hash = update_query_hash(record)
+        print(f"Updating hash for query {record['query_id']} from {record['query_hash']} to {new_hash}")
+        conn.execute(
+            queries.update()
+            .where(queries.c.id == record['query_id'])
+            .values(query_hash=new_hash))
+
+
+def downgrade():
+    pass

redash/cli/queries.py

@@ -5,6 +5,22 @@
 manager = AppGroup(help="Queries management commands.")
 
 
+@manager.command(name="rehash")
+def rehash():
+    from redash import models
+
+    for q in models.Query.query.all():
+        old_hash = q.query_hash
+        q.update_query_hash()
+        new_hash = q.query_hash
+
+        if old_hash != new_hash:
+            print(f"Query {q.id} has changed hash from {old_hash} to {new_hash}")
+            models.db.session.add(q)
+
+    models.db.session.commit()
+
+
 @manager.command(name="add_tag")
 @argument("query_id")
 @argument("tag")

redash/metrics/database.py

@@ -5,7 +5,7 @@
 from sqlalchemy.engine import Engine
 from sqlalchemy.event import listens_for
 from sqlalchemy.orm.util import _ORMJoin
-from sqlalchemy.sql.selectable import Alias
+from sqlalchemy.sql.selectable import Alias, Join
 
 from redash import statsd_client
 
@@ -18,7 +18,7 @@ def _table_name_from_select_element(elt):
     if isinstance(t, Alias):
         t = t.original.froms[0]
 
-    while isinstance(t, _ORMJoin):
+    while isinstance(t, _ORMJoin) or isinstance(t, Join):
         t = t.left
 
     return t.name