group_concat 을 이용하면됩니다,

조회해보세요

 괄호안은 이해를 돕기 위해 표시한 정보입니다. :)

우선 다른 사이트에서 위 쿼리를 해결하는 방법을 알려주셨습니다.

CREATE TABLE `test1` (

  `date_` varchar(8) DEFAULT NULL,
  `user_id` varchar(5) DEFAULT NULL,
  `gold` int(11) DEFAULT NULL
)

SELECT z.inputdate DATE, SUM(gold) sum_of_amount FROM test1 a, (
SELECT inputdate, user_id, MAX(date_) mydate FROM (
SELECT  b.date_ inputdate, a.user_id, a.date_, a.gold  FROM test1 a, (SELECT DISTINCT(date_) FROM test1 ORDER BY date_ ASC) b WHERE A.DATE_<= B.DATE_ ORDER BY b.date_, a.user_id ASC
) Y  GROUP BY inputdate, user_id
) z WHERE a.user_id = z.user_id AND a.date_ = z.mydate

하지만 데이터가 워낙 방대하여 성능상의 이유로 결국은 다른 방법을 찾아야 하네요.ㅠ

SELECT date_
     , @x := @x + gold AS gold
  FROM (SELECT date_
             , SUM(gold) gold
          FROM (SELECT user_id, date_
                     , gold - CASE @user_id WHEN user_id
                              THEN @gold ELSE 0 END AS gold
                     , @gold    := gold
                     , @user_id := user_id
                  FROM (SELECT user_id, date_, gold
                          FROM test1
                         ORDER BY user_id, date_
                        ) a
                     , (SELECT @user_id := '', @gold := 0, @x := 0) b
                ) a
         GROUP BY date_
        ) a
;