diff --git a/scripts/jobs/rentsense_former_tenants_to_refined.py b/scripts/jobs/rentsense_former_tenants_to_refined.py
index ac258ba8e..1964e5a19 100644
--- a/scripts/jobs/rentsense_former_tenants_to_refined.py
+++ b/scripts/jobs/rentsense_former_tenants_to_refined.py
@@ -23,6 +23,7 @@
     datediff,
     when,
 )
+from pyspark.sql.window import Window
 from scripts.helpers.helpers import (
     PARTITION_KEYS,
     add_import_time_columns,
@@ -980,23 +981,20 @@ def export_dynamic_frame_as_xml_gzip(
         case_priorities, accounts4.tenancy_ref == case_priorities.tenancy_ref2, "left"
     )
 
-    accounts6 = accounts5.selectExpr(
-        "AccountReference as AccountReference",
-        # "TenureType",
-        #  "TenureTypeCode",
-        # "max_date as TenancyStartDate",
-        "TenancyEndDate",
-        "LocalAuthority",
-        # "HousingOfficerName",
-        "Patch",
-        "'Hackney' as Region",
-        #   "import_date as import_date",
-        #  "tenancy_ref as TenReference",
-        #  "is_paused_until as BreathingSpaceEndDate",
-        "Case when Deceased=1 then 'Y' else 'N' end as Deceased"
-        #  "previousweekbalance"
-    )
+    w = Window.partitionBy("AccountReference").orderBy(F.col("TenancyEndDate").desc())
 
+    accounts6 = (
+        accounts5.withColumn("rn", F.row_number().over(w))
+        .filter(F.col("rn") == 1)
+        .selectExpr(
+            "AccountReference",
+            "TenancyEndDate",
+            "LocalAuthority",
+            "Patch",
+            "'Hackney' as Region",
+            "case when Deceased=1 then 'Y' else 'N' end as Deceased",
+        )
+    )
     accounts7 = accounts6.filter("AccountReference is not null")
 
     accounts8 = accounts7.distinct()