[CLIENT-3972] Dev tests: paginated query test case should handle receiving less than all the records when running against a multi-node cluster and the records are unevenly distributed across the nodes / partitions (#992)

Author: juliannguyen4

test/new_tests/test_query_pagination.py

@@ -130,25 +130,28 @@ def callback(part_id, input_tuple):
         query_obj = self.as_connection.query(self.test_ns, None)
         query_obj.paginate()
 
-        num_populated_partitions = 4
-        all_records = (
+        NUM_PARTITIONS = 4
+        num_records_from_part_1000_to_1003 = (
             self.partition_1000_count
             + self.partition_1001_count
             + self.partition_1002_count
             + self.partition_1003_count
         )
-        self.partition_1000_count / num_populated_partitions
-        query_obj.max_records = math.ceil(all_records / num_populated_partitions)
+        avg_records_per_partition = math.ceil(num_records_from_part_1000_to_1003 / NUM_PARTITIONS)
+        query_obj.max_records = avg_records_per_partition
 
-        for i in range(num_populated_partitions):
+        NUM_ITERATIONS = NUM_PARTITIONS
+        for _ in range(NUM_ITERATIONS):
             query_obj.foreach(
                 callback,
                 {
-                    "partition_filter": {"begin": 1000, "count": num_populated_partitions},
+                    "partition_filter": {"begin": 1000, "count": NUM_PARTITIONS},
                 },
             )
 
-        assert len(records) == all_records
+        # Worst case scenario, all the records are in one node
+        # Best case scenario, we got all the records back
+        assert NUM_ITERATIONS <= len(records) <= num_records_from_part_1000_to_1003
 
     # NOTE: This could fail if node record counts are small and unbalanced across nodes.
     @pytest.mark.xfail(reason="Might fail depending on record count and distribution.")