С ROLLUP в сочетании с критериями MULPLIPLE GROUP BY

Question

У меня есть следующая таблица:

CREATE TABLE PaL (
    Event_Date DATE,
    Country CHAR(2),
    Category CHAR(255),
    Revenue INTEGER(255),
    Costs INTEGER(255)
);

INSERT INTO PaL 
(Event_Date, Country, Category, Revenue, Costs)
VALUES

("2017-01-31", "DE", "Apparel", "692.09816652375", "-173.071989376023"),
("2017-02-28", "DE", "Apparel", "8419.9977988914", "-7622.61265984317"),
("2017-03-31", "DE", "Apparel", "2018.80471444031", "-1498.76213884283"),
("2017-04-30", "DE", "Apparel", "8863.15663035884", "-7965.69268589649"),
("2017-05-31", "DE", "Apparel", "6838.4514829573", "-1088.70351845663"),
("2017-06-30", "DE", "Apparel", "2025.73421386331", "-483.454199185678"),
("2017-07-31", "DE", "Apparel", "5389.0163788639", "-2643.93624645182"),
("2017-08-31", "DE", "Apparel", "6238.85870250446", "-1985.9879371866"),
("2017-09-30", "DE", "Apparel", "2294.62451106469", "-1864.98271539745"),
("2017-10-31", "DE", "Apparel", "4141.2074159951", "-197.773961036073"),
("2017-11-30", "DE", "Apparel", "1456.17584217397", "-1018.54129047119"),
("2017-12-31", "DE", "Apparel", "3623.54984724091", "-745.715567286581"),

("2017-01-31", "DE", "Shoes", "5955.20947079185", "-4745.39564508682"),
("2017-02-28", "DE", "Shoes", "9555.29563511224", "-5729.82601329738"),
("2017-03-31", "DE", "Shoes", "5490.36170257556", "-925.286457266534"),
("2017-04-30", "DE", "Shoes", "7652.35548686073", "-7335.32532050594"),
("2017-05-31", "DE", "Shoes", "9102.38987703511", "-5724.92574170071"),
("2017-06-30", "DE", "Shoes", "1703.95901703023", "-1678.19547060803"),
("2017-07-31", "DE", "Shoes", "3679.60045104324", "-2095.94207835501"),
("2017-08-31", "DE", "Shoes", "6672.43210841331", "-3475.55411014914"),
("2017-09-30", "DE", "Shoes", "7718.7744220635", "-1252.75877307055"),
("2017-10-31", "DE", "Shoes", "6976.01564153854", "-509.991595559256"),
("2017-11-30", "DE", "Shoes", "4725.46976319905", "-2835.09460170927"),
("2017-12-31", "DE", "Shoes", "8390.84483147949", "-7476.83516162742"),


("2017-01-31", "US", "Apparel", "939788.159047677", "-742666.846083707"),
("2017-02-28", "US", "Apparel", "826418.514009279", "-702997.151099908"),
("2017-03-31", "US", "Apparel", "775940.69563018", "-211238.971709086"),
("2017-04-30", "US", "Apparel", "516829.583069596", "-407521.856789393"),
("2017-05-31", "US", "Apparel", "635701.377748304", "-627829.016481388"),
("2017-06-30", "US", "Apparel", "757852.95599751", "-740948.867522139"),
("2017-07-31", "US", "Apparel", "154224.257732688", "-139805.456987081"),
("2017-08-31", "US", "Apparel", "102035.465731255", "-100103.875992667"),
("2017-09-30", "US", "Apparel", "880671.692714021", "-665324.083753931"),
("2017-10-31", "US", "Apparel", "187868.653562564", "-105676.793254039"),
("2017-11-30", "US", "Apparel", "994600.486892401", "-177382.899789215"),
("2017-12-31", "US", "Apparel", "813824.90461202", "-132527.311010471"),

("2017-01-31", "US", "Shoes", "661069.933966637", "-454778.427240679"),
("2017-02-28", "US", "Shoes", "675942.334464692", "-254489.623313569"),
("2017-03-31", "US", "Shoes", "473604.307973888", "-404226.047653847"),
("2017-04-30", "US", "Shoes", "872018.822577053", "-348781.396359871"),
("2017-05-31", "US", "Shoes", "718012.023481434", "-625306.312927362"),
("2017-06-30", "US", "Shoes", "688487.679029354", "-584512.575888519"),
("2017-07-31", "US", "Shoes", "690085.013711018", "-581753.771085971"),
("2017-08-31", "US", "Shoes", "204473.88894161", "-172301.871771595"),
("2017-09-30", "US", "Shoes", "516932.092423463", "-328002.737710081"),
("2017-10-31", "US", "Shoes", "609355.245817292", "-323624.391366703"),
("2017-11-30", "US", "Shoes", "313599.625504231", "-210253.020497022"),
("2017-12-31", "US", "Shoes", "723573.681040319", "-107333.764977439"),


("2017-01-31", "NZ", "Apparel", "81292.9610624533", "-27354.678748396"),
("2017-02-28", "NZ", "Apparel", "77473.6231986387", "-47920.2900396812"),
("2017-03-31", "NZ", "Apparel", "93819.4377266116", "-47582.1878235771"),
("2017-04-30", "NZ", "Apparel", "25580.8516093492", "-21277.2651303701"),
("2017-05-31", "NZ", "Apparel", "82842.9415935231", "-30714.5952859941"),
("2017-06-30", "NZ", "Apparel", "50878.6190715448", "-33047.3841488076"),
("2017-07-31", "NZ", "Apparel", "61423.3558286064", "-10811.2817583225"),
("2017-08-31", "NZ", "Apparel", "77517.2989019148", "-56818.7461336424"),
("2017-09-30", "NZ", "Apparel", "74046.1258000888", "-10108.0702908427"),
("2017-10-31", "NZ", "Apparel", "79490.650598675", "-68562.5753547413"),
("2017-11-30", "NZ", "Apparel", "65000.3971251097", "-25174.1329786955"),
("2017-12-31", "NZ", "Apparel", "99152.6457285608", "-42855.8431883814"),

("2017-01-31", "NZ", "Shoes", "20703.8970205884", "-11911.9616025915"),
("2017-02-28", "NZ", "Shoes", "72841.2537140946", "-14166.6747335237"),
("2017-03-31", "NZ", "Shoes", "45391.6550622383", "-40325.1638601903"),
("2017-04-30", "NZ", "Shoes", "58074.2843201579", "-54483.1122507654"),
("2017-05-31", "NZ", "Shoes", "52127.2701338519", "-28026.7984458694"),
("2017-06-30", "NZ", "Shoes", "32900.9222204099", "-22780.2637095601"),
("2017-07-31", "NZ", "Shoes", "18809.3868235169", "-11500.4020522949"),
("2017-08-31", "NZ", "Shoes", "67001.2729206886", "-53280.8129552599"),
("2017-09-30", "NZ", "Shoes", "26889.4058005421", "-24218.8734875798"),
("2017-10-31", "NZ", "Shoes", "56330.7544011198", "-51382.4201254223"),
("2017-11-30", "NZ", "Shoes", "60954.7129549264", "-19834.7256352672"),
("2017-12-31", "NZ", "Shoes", "97527.2014993995", "-83137.4844853141");

И я использую следующий запрос для получения данных из таблицы:

Select Country, Category, sum(Revenue) as Revenue, sum(Costs) as Costs
FROM Pal
WHERE Event_Date BETWEEN "2017-01-01" AND "2017-01-31"
GROUP BY Country, Category WITH ROLLUP

Вы также можете найти таблицу с данными в sql fiddle здесь
Пока все это прекрасно работает.

---

Теперь мне было интересно, как можно избежать, чтобы функция WITH ROLLUP вычисляла сумму столбца под каждой страной. Вместо этого он должен вычислять итоговое значение столбца только один раз, поэтому результат в итоге будет выглядеть следующим образом:

Country       Category      Revenue     Costs
DE              Apparel      692         -173
DE              Shoes        5955        -4745
:                 :           :            :
:                 :           :            :
:                 :           :            :
US              Shoes        661070       -454778
(null)         (null)       1709502      -1241630

Что мне нужно изменить в моем запросе SQL, чтобы добиться этого?

Answer 1

Удалить с накоплением

Select Country, Category, sum(Revenue) as Revenue, sum(Costs) as Costs
FROM Pal
WHERE Event_Date BETWEEN "2017-01-01" AND "2017-01-31"
GROUP BY Country, Category 

и затем используйте union all, как сэр @Gordon использует свой ответ

Answer 2

Вы можете использовать HAVING, чтобы отфильтровать промежуточные итоги для каждой страны:

Select Country, Category, sum(Revenue) as Revenue, sum(Costs) as Costs
FROM Pal
WHERE Event_Date BETWEEN "2017-01-01" AND "2017-01-31"
GROUP BY Country, Category WITH ROLLUP
HAVING (Country IS NULL AND Category IS NULL) OR (Country IS NOT NULL AND Category IS NOT NULL)

Условие Country IS NULL AND Category IS NULL соответствует общему итогу в конце, условие Country IS NOT NULL AND Category IS NOT NULL соответствует отдельным строкам для каждой страны и категории.

DEMO

Answer 3

MySQL не поддерживает GROUPING SETS, что вам действительно нужно. Пожалуй, самый простой способ - использовать UNION ALL:

SELECT Country, Category, SUM(Revenue) as Revenue, SUM(Costs) as Costs
FROM Pal
WHERE Event_Date BETWEEN '2017-01-01' AND '2017-01-31'
GROUP BY Country, Category 
UNION ALL
SELECT NULL, NULL, SUM(Revenue) as Revenue, SUM(Costs) as Costs
FROM Pal
WHERE Event_Date BETWEEN '2017-01-01' AND '2017-01-31';