diff --git a/erpnext/buying/report/procurement_tracker/procurement_tracker.py b/erpnext/buying/report/procurement_tracker/procurement_tracker.py
index 10169c554fb..fd30ddf6884 100644
--- a/erpnext/buying/report/procurement_tracker/procurement_tracker.py
+++ b/erpnext/buying/report/procurement_tracker/procurement_tracker.py
@@ -305,7 +305,9 @@ def get_po_entries(filters):
 			& (parent.name == child.parent)
 			& (parent.status.notin(("Closed", "Completed", "Cancelled")))
 		)
-		.groupby(parent.name, child.material_request_item)
+		# This is one row per PO item; the selected child.* columns are only functionally dependent
+		# on the child PK, which postgres requires in the GROUP BY (MariaDB allows omitting it).
+		.groupby(parent.name, child.material_request_item, child.name)
 	)
 	query = apply_filters_on_query(filters, parent, child, query)
 
diff --git a/erpnext/buying/report/procurement_tracker/test_procurement_tracker.py b/erpnext/buying/report/procurement_tracker/test_procurement_tracker.py
index 1577bf2cbf6..f1781c15e05 100644
--- a/erpnext/buying/report/procurement_tracker/test_procurement_tracker.py
+++ b/erpnext/buying/report/procurement_tracker/test_procurement_tracker.py
@@ -6,4 +6,16 @@ from erpnext.tests.utils import ERPNextTestSuite
 
 
 class TestProcurementTracker(ERPNextTestSuite):
-	pass
+	def test_report_executes_and_lists_po(self):
+		# get_po_entries groups by (Purchase Order, material_request_item, Purchase Order Item)
+		# while selecting other child columns; this exercises that GROUP BY so the report stays
+		# valid on Postgres (which rejects selecting non-grouped columns).
+		from erpnext.buying.doctype.purchase_order.test_purchase_order import create_purchase_order
+		from erpnext.buying.report.procurement_tracker.procurement_tracker import execute
+
+		po = create_purchase_order(company="_Test Company")
+
+		columns, data = execute({"company": "_Test Company"})
+
+		self.assertTrue(columns)
+		self.assertIn(po.name, {row.get("purchase_order") for row in data})