MySQL WITH ROLLUP when grouping by multiple fields

Question

Performing a WITH ROLLUP when grouping by multiple fields, MySQL returns a rollup row for each group. I'm interested in the rollups for fields number and perc in dotable below:

+------+---------------+--------+-------+--------+------------+
| RDT  | type          | number | total | perc   | thedate    |
+------+---------------+--------+-------+--------+------------+
| MAL  | checking      |    112 |  3249 | 3.4    | 2014-03-27 |
| MAL  | control       |     33 |  3249 | 1.0    | 2014-03-27 |
| MAL  | reconstructed |      5 |  3249 | 0.2    | 2014-03-27 |
| MAL  | regular       |    960 |  3249 | 29.5   | 2014-03-27 |
| MAL  | study         |     10 |  3249 | 0.3    | 2014-03-27 |
| tot  | NULL          |   1120 |  3249 | 34.0   | 2014-03-27 |

| PMC  | checking      |    107 |  3153 | 3.4    | 2014-03-27 |
| PMC  | control       |    167 |  3153 | 5.3    | 2014-03-27 |
| PMC  | reconstructed |      8 |  3153 | 0.3    | 2014-03-27 |
| PMC  | regular       |    833 |  3153 | 26.4   | 2014-03-27 |
| PMC  | study         |     72 |  3153 | 2.3    | 2014-03-27 |
| tot  | NULL          |   1187 |  3249 | 36.5   | 2014-03-27 |

| RAS  | checking      |      1 |   970 | 0.1    | 2014-03-27 |
| RAS  | control       |     42 |   970 | 4.3    | 2014-03-27 |
| RAS  | reconstructed |      1 |   970 | 0.1    | 2014-03-27 |
| RAS  | regular       |    318 |   970 | 32.8   | 2014-03-27 |
| RAS  | study         |      3 |   970 | 0.3    | 2014-03-27 |
| tot  | NULL          |    365 |   970 | 37.6   | 2014-03-27 |

| UOT  | checking      |     11 |  3527 | 0.3    | 2014-03-27 |
| UOT  | control       |    283 |  3527 | 8.0    | 2014-03-27 |
| UOT  | regular       |    235 |  3527 | 6.7    | 2014-03-27 |
| UOT  | study         |      8 |  3527 | 0.2    | 2014-03-27 |
| tot  | NULL          |    537 |  3527 | 15.2   | 2014-03-27 |
+------+---------------+--------+-------+--------+------------+

And I tried this query:

mysql> SELECT
    IFNULL(`RDT`, 'tot') AS RDT,
    `type`,
    `number`,
    `total`,
    `perc`,
    `thedate`
FROM
    (
        SELECT
            `RDT`,
            `type`,
            `NUMBER`,
            `total`,
            `perc`,
            `THEDATE`
        FROM
            `dotable`
        GROUP BY
            RDT,
            `type`
        ORDER BY
            CASE
        WHEN RDT = 'UOT' THEN
            1
        END ASC
    ) AS X
GROUP BY
    X.RDT,
    X.`type` WITH ROLLUP;

But the output is wrong, can you help me? the dotable structure and rows is below.

+------+---------------+--------+-------+--------+------------+
| RDT  | type          | number | total | perc   | thedate    |
+------+---------------+--------+-------+--------+------------+
| MAL  | checking      |    112 |  3249 | 3.4    | 2014-03-27 |
| MAL  | control       |     33 |  3249 | 1.0    | 2014-03-27 |
| MAL  | reconstructed |      5 |  3249 | 0.2    | 2014-03-27 |
| MAL  | regular       |    960 |  3249 | 29.5   | 2014-03-27 |
| MAL  | study         |     10 |  3249 | 0.3    | 2014-03-27 |
| MAL  | NULL          |     10 |  3249 | 0.3    | 2014-03-27 |
| PMC  | checking      |    107 |  3153 | 3.4    | 2014-03-27 |
| PMC  | control       |    167 |  3153 | 5.3    | 2014-03-27 |
| PMC  | reconstructed |      8 |  3153 | 0.3    | 2014-03-27 |
| PMC  | regular       |    833 |  3153 | 26.4   | 2014-03-27 |
| PMC  | study         |     72 |  3153 | 2.3    | 2014-03-27 |
| PMC  | NULL          |     72 |  3153 | 2.3    | 2014-03-27 |
| RAS  | checking      |      1 |   970 | 0.1    | 2014-03-27 |
| RAS  | control       |     42 |   970 | 4.3    | 2014-03-27 |
| RAS  | reconstructed |      1 |   970 | 0.1    | 2014-03-27 |
| RAS  | regular       |    318 |   970 | 32.8   | 2014-03-27 |
| RAS  | study         |      3 |   970 | 0.3    | 2014-03-27 |
| RAS  | NULL          |      3 |   970 | 0.3    | 2014-03-27 |
| UOT  | checking      |     11 |  3527 | 0.3    | 2014-03-27 |
| UOT  | control       |    283 |  3527 | 8.0    | 2014-03-27 |
| UOT  | regular       |    235 |  3527 | 6.7    | 2014-03-27 |
| UOT  | study         |      8 |  3527 | 0.2    | 2014-03-27 |
| UOT  | NULL          |      8 |  3527 | 0.2    | 2014-03-27 |
| tot  | NULL          |      8 |  3527 | 0.2    | 2014-03-27 |
+------+---------------+--------+-------+--------+------------+
24 rows in set

mysql> 

SET FOREIGN_KEY_CHECKS=0;

-- ----------------------------
-- Table structure for `dotable`
-- ----------------------------
DROP TABLE IF EXISTS `dotable`;
CREATE TABLE `dotable` (
  `RDT` varchar(10) DEFAULT NULL,
  `TYPE` varchar(50) DEFAULT NULL,
  `NUMBER` int(11) NOT NULL DEFAULT '0',
  `TOTAL` int(11) NOT NULL DEFAULT '0',
  `PERC` decimal(10,2) DEFAULT NULL,
  `THEDATE` date DEFAULT NULL,
  `ID` int(10) NOT NULL AUTO_INCREMENT,
  PRIMARY KEY (`ID`)
) ENGINE=MyISAM AUTO_INCREMENT=20 DEFAULT CHARSET=latin1;

-- ----------------------------
-- Records of dotable
-- ----------------------------
INSERT INTO `dotable` VALUES ('UOT', 'control', '283', '3527', '8.0', '2014-03-27', '1');
INSERT INTO `dotable` VALUES ('UOT', 'regular', '235', '3527', '6.7', '2014-03-27', '2');
INSERT INTO `dotable` VALUES ('UOT', 'study', '8', '3527', '0.2', '2014-03-27', '3');
INSERT INTO `dotable` VALUES ('UOT', 'checking', '11', '3527', '0.3', '2014-03-27', '4');
INSERT INTO `dotable` VALUES ('MAL', 'regular', '960', '3249', '29.5', '2014-03-27', '5');
INSERT INTO `dotable` VALUES ('MAL', 'study', '10', '3249', '0.3', '2014-03-27', '6');
INSERT INTO `dotable` VALUES ('MAL', 'control', '33', '3249', '1.0', '2014-03-27', '7');
INSERT INTO `dotable` VALUES ('MAL', 'reconstructed', '5', '3249', '0.2', '2014-03-27', '8');
INSERT INTO `dotable` VALUES ('MAL', 'checking', '112', '3249', '3.4', '2014-03-27', '9');
INSERT INTO `dotable` VALUES ('PMC', 'regular', '833', '3153', '26.4', '2014-03-27', '10');
INSERT INTO `dotable` VALUES ('PMC', 'study', '72', '3153', '2.3', '2014-03-27', '11');
INSERT INTO `dotable` VALUES ('PMC', 'control', '167', '3153', '5.3', '2014-03-27', '12');
INSERT INTO `dotable` VALUES ('PMC', 'checking', '107', '3153', '3.4', '2014-03-27', '13');
INSERT INTO `dotable` VALUES ('PMC', 'reconstructed', '8', '3153', '0.3', '2014-03-27', '14');
INSERT INTO `dotable` VALUES ('RAS', 'reconstructed', '1', '970', '0.1', '2014-03-27', '15');
INSERT INTO `dotable` VALUES ('RAS', 'regular', '318', '970', '32.8', '2014-03-27', '16');
INSERT INTO `dotable` VALUES ('RAS', 'study', '3', '970', '0.3', '2014-03-27', '17');
INSERT INTO `dotable` VALUES ('RAS', 'control', '42', '970', '4.3', '2014-03-27', '18');
INSERT INTO `dotable` VALUES ('RAS', 'checking', '1', '970', '0.1', '2014-03-27', '19');

Answer 1

I'm interested in the rollups for fields number and perc in dotable.

So you need to do a sum() to make it work.

If you don't need the overall total we can also remove it.

For Version <= 5.7

Query:

select 
    if(`type` is null, 'tot', `rdt`) as `rdt`,
    `type`,
    `number`,
    `total`,
    `perc`,
    `thedate`
from
    (select 
        `rdt`,
        `type`,
        sum(`number`) as `number`,
        `total`,
        sum(`perc`) as `perc`,
        `thedate`
    from
        `dotable`
    group by `rdt`, `type` with rollup) as t
where
    `rdt` is not null;

Result:

rdt | type          | number | total |  perc | thedate   
:-- | :------------ | -----: | ----: | ----: | :---------
MAL | checking      |    112 |  3249 |  3.40 | 2014-03-27
MAL | control       |     33 |  3249 |  1.00 | 2014-03-27
MAL | reconstructed |      5 |  3249 |  0.20 | 2014-03-27
MAL | regular       |    960 |  3249 | 29.50 | 2014-03-27
MAL | study         |     10 |  3249 |  0.30 | 2014-03-27
tot | null          |   1120 |  3249 | 34.40 | 2014-03-27
PMC | checking      |    107 |  3153 |  3.40 | 2014-03-27
PMC | control       |    167 |  3153 |  5.30 | 2014-03-27
PMC | reconstructed |      8 |  3153 |  0.30 | 2014-03-27
PMC | regular       |    833 |  3153 | 26.40 | 2014-03-27
PMC | study         |     72 |  3153 |  2.30 | 2014-03-27
tot | null          |   1187 |  3153 | 37.70 | 2014-03-27
RAS | checking      |      1 |   970 |  0.10 | 2014-03-27
RAS | control       |     42 |   970 |  4.30 | 2014-03-27
RAS | reconstructed |      1 |   970 |  0.10 | 2014-03-27
RAS | regular       |    318 |   970 | 32.80 | 2014-03-27
RAS | study         |      3 |   970 |  0.30 | 2014-03-27
tot | null          |    365 |   970 | 37.60 | 2014-03-27
UOT | checking      |     11 |  3527 |  0.30 | 2014-03-27
UOT | control       |    283 |  3527 |  8.00 | 2014-03-27
UOT | regular       |    235 |  3527 |  6.70 | 2014-03-27
UOT | study         |      8 |  3527 |  0.20 | 2014-03-27
tot | null          |    537 |  3527 | 15.20 | 2014-03-27

---

For Version > 5.7

Query:

select 
    if(`type` is null, 'tot', `rdt`) as `rdt`,
    `type`,
    sum(`number`) as `number`,
    `total`,
    sum(`perc`) as `perc`,
    `thedate`
from
    `dotable`
group by `rdt`, `type` with rollup
having not grouping(rdt) <> 0;

Result:

rdt | type          | number | total |  perc | thedate   
:-- | :------------ | -----: | ----: | ----: | :---------
MAL | checking      |    112 |  3249 |  3.40 | 2014-03-27
MAL | control       |     33 |  3249 |  1.00 | 2014-03-27
MAL | reconstructed |      5 |  3249 |  0.20 | 2014-03-27
MAL | regular       |    960 |  3249 | 29.50 | 2014-03-27
MAL | study         |     10 |  3249 |  0.30 | 2014-03-27
tot | null          |   1120 |  3249 | 34.40 | 2014-03-27
PMC | checking      |    107 |  3153 |  3.40 | 2014-03-27
PMC | control       |    167 |  3153 |  5.30 | 2014-03-27
PMC | reconstructed |      8 |  3153 |  0.30 | 2014-03-27
PMC | regular       |    833 |  3153 | 26.40 | 2014-03-27
PMC | study         |     72 |  3153 |  2.30 | 2014-03-27
tot | null          |   1187 |  3153 | 37.70 | 2014-03-27
RAS | checking      |      1 |   970 |  0.10 | 2014-03-27
RAS | control       |     42 |   970 |  4.30 | 2014-03-27
RAS | reconstructed |      1 |   970 |  0.10 | 2014-03-27
RAS | regular       |    318 |   970 | 32.80 | 2014-03-27
RAS | study         |      3 |   970 |  0.30 | 2014-03-27
tot | null          |    365 |   970 | 37.60 | 2014-03-27
UOT | checking      |     11 |  3527 |  0.30 | 2014-03-27
UOT | control       |    283 |  3527 |  8.00 | 2014-03-27
UOT | regular       |    235 |  3527 |  6.70 | 2014-03-27
UOT | study         |      8 |  3527 |  0.20 | 2014-03-27
tot | null          |    537 |  3527 | 15.20 | 2014-03-27

---

Fiddle:
db<>fiddle here

Note: When using group by without ERROR 1055 you need to set as

SET sql_mode=(SELECT REPLACE(@@sql_mode,'ONLY_FULL_GROUP_BY',''));