fix: update DISTINCT query to avoid database context issues

Author: tae898

bindings/python/examples/05_csv_import_graph.py

@@ -300,7 +300,9 @@ def create_all_vertices(
     def _create_users(self, total_users: int) -> tuple[int, BenchmarkStats]:
         """Create User vertices.
 
-        Note: Uses SELECT DISTINCT (not paginated - efficient pagination difficult)
+        Note: Uses a DISTINCT subquery. The direct DISTINCT/ORDER BY form can
+        resolve against the wrong database context when the target graph DB is
+        open at the same time as the source document DB.
         """
         print(f"Creating {total_users:,} User vertices...")
         stats = BenchmarkStats()
@@ -317,7 +319,10 @@ def _create_users(self, total_users: int) -> tuple[int, BenchmarkStats]:
             with arcadedb.open_database(
                 str(self.data_loader.source_db_path)
             ) as source_db:
-                query = "SELECT DISTINCT userId FROM Rating ORDER BY userId"
+                query = (
+                    "SELECT userId FROM (SELECT DISTINCT userId FROM Rating) "
+                    "ORDER BY userId"
+                )
                 for record in source_db.query("sql", query):
                     user_id = record.get("userId")
                     async_exec.command(
@@ -344,7 +349,10 @@ def _create_users(self, total_users: int) -> tuple[int, BenchmarkStats]:
             with arcadedb.open_database(
                 str(self.data_loader.source_db_path)
             ) as source_db:
-                query = "SELECT DISTINCT userId FROM Rating ORDER BY userId"
+                query = (
+                    "SELECT userId FROM (SELECT DISTINCT userId FROM Rating) "
+                    "ORDER BY userId"
+                )
                 batch_user_ids = []
 
                 for record in source_db.query("sql", query):
@@ -380,7 +388,10 @@ def _create_users(self, total_users: int) -> tuple[int, BenchmarkStats]:
             with arcadedb.open_database(
                 str(self.data_loader.source_db_path)
             ) as source_db:
-                query = "SELECT DISTINCT userId FROM Rating ORDER BY userId"
+                query = (
+                    "SELECT userId FROM (SELECT DISTINCT userId FROM Rating) "
+                    "ORDER BY userId"
+                )
                 batch_user_ids = []
 
                 for record in source_db.query("sql", query):