Use unique dataset ids when testing routines (#1416)

Co-authored-by: Frank Bertsch <fbertsch@mozilla.com>

bigquery_etl/pytest_plugin/routine.py

@@ -2,6 +2,7 @@
 
 from google.api_core.exceptions import BadRequest
 from google.cloud import bigquery
+import os
 import pytest
 import re
 
@@ -91,7 +92,10 @@ class RoutineTest(pytest.Item):
     def runtest(self):
         """Run Test."""
         bq = bigquery.Client()
-        with dataset(bq, self.safe_name()) as default_dataset:
+        dataset_id = self.safe_name()
+        if "CIRCLE_BUILD_NUM" in os.environ:
+            dataset_id += f"_{os.environ['CIRCLE_BUILD_NUM']}"
+        with dataset(bq, dataset_id) as default_dataset:
             job_config = bigquery.QueryJobConfig(
                 use_legacy_sql=False, default_dataset=default_dataset
             )