From 2dcf2af2a0d0ff1fe4f415cf69d63a35ea2c66f4 Mon Sep 17 00:00:00 2001
From: Alex Parsons <alex.parsons@mysociety.org>
Date: Fri, 11 Aug 2023 08:15:21 +0000
Subject: [PATCH] Remove __index_level_0__ from csv exports

---
 src/data_common/dataset/resource_management.py | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/src/data_common/dataset/resource_management.py b/src/data_common/dataset/resource_management.py
index 1ec53ce..f504df6 100644
--- a/src/data_common/dataset/resource_management.py
+++ b/src/data_common/dataset/resource_management.py
@@ -918,9 +918,12 @@ def copy_resources(self):
         csv_copy_query = """
         copy (select * from {{ source }}) to {{ dest }} (format PARQUET);
         """
-        exclude = ""
+
+        # __index_level_0__ is an internal parquet column that duckdb has access to
+        # but we don't want to export
+        exclude = "EXCLUDE __index_level_0__"
         if desc["custom"].get("is_geodata", False):
-            exclude = "EXCLUDE geometry"
+            exclude = "EXCLUDE __index_level_0__, geometry"
 
         parquet_copy_query = """
         copy (select * {{ exclude }} from {{ source }}) to {{ dest }} (HEADER, DELIMITER ',');