Modify SELECT to correct situation when data is missing from MyRocks table when GROUP BY is used

CREATE TABLE `test` (
`col1` bigint unsigned NOT NULL AUTO_INCREMENT,
`col2` binary(16) NOT NULL,
`col3` varchar(191) CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci NOT NULL,
`col4` bigint NOT NULL,
`col5` mediumblob NOT NULL,
PRIMARY KEY (`col1`),
UNIQUE KEY `uc_key` (`col2`,`col3`,`col4`)
) ENGINE=ROCKSDB;
insert into test (col2,col3,col4,col5) values('aa', 'aa', 100, 'aa');
insert into test (col2,col3,col4,col5) values('aa', 'bb', 100, 'bb');
insert into test (col2,col3,col4,col5) values('aa', 'cc', 100, 'cc');
insert into test (col2,col3,col4,col5) values('aa', 'dd', 100, 'dd');
insert into test (col2,col3,col4,col5) values('aa', 'cc', 99, 'cc');

SELECT col2, col3, max(col4) AS col4 FROM test t WHERE t.col2 IN (x'61610000000000000000000000000000') AND t.col3 IN ('aa','bb','cc','dd') GROUP BY t.col2, t.col3;

+------------------------------------+------+------+
| col2 | col3 | col4 |
+------------------------------------+------+------+
| 0x61610000000000000000000000000000 | aa | 100 |
| 0x61610000000000000000000000000000 | bb | 100 |
| 0x61610000000000000000000000000000 | cc | 100 |
| 0x61610000000000000000000000000000 | dd | 100 |
+------------------------------------+------+------+
4 rows in set (0.00 sec)

alter table test engine=rocksdb;

+------------------------------------+------+------+
| col2 | col3 | col4 |
+------------------------------------+------+------+
| 0x61610000000000000000000000000000 | aa | 100 |
+------------------------------------+------+------+
1 row in set (0.00 sec)

--- explain.good 2020-09-16 15:16:39.410862546 -0300
+++ explain.bad 2020-09-16 15:14:59.152933716 -0300
@@ -2,7 +2,7 @@
 "query_block": {
 "select_id": 1,
 "cost_info": {
- "query_cost": "1.35"
+ "query_cost": "2.00"
 },
 "grouping_operation": {
 "using_filesort": false,
@@ -18,15 +18,15 @@
 "col3"
 ],
 "key_length": "782",
- "rows_examined_per_scan": 5,
- "rows_produced_per_join": 5,
+ "rows_examined_per_scan": 7,
+ "rows_produced_per_join": 7,
 "filtered": "100.00",
- "using_index": true,
+ "using_index_for_group_by": true,
 "cost_info": {
- "read_cost": "0.85",
- "eval_cost": "0.50",
- "prefix_cost": "1.35",
- "data_read_per_join": "3K"
+ "read_cost": "1.30",
+ "eval_cost": "0.70",
+ "prefix_cost": "2.00",
+ "data_read_per_join": "5K"
 },
 "used_columns": [
 "col1",

SELECT /*+ NO_RANGE_OPTIMIZATION(t uc_key)*/ col2, col3, max(col4) AS col4 FROM test t WHERE t.col2 IN (x'61610000000000000000000000000000') AND t.col3 IN ('aa','bb','cc','dd') GROUP BY t.col2, t.col3;