WITH t AS
(
SELECT '2016-01-01 01' reg_dt, '홍길동' id, '로그인' gb, 100 milg FROM dual
UNION ALL SELECT '2016-01-01 02', '홍길동', '글등록'  , 1000 FROM dual
UNION ALL SELECT '2016-01-01 03', '홍길동', '댓글등록',  500 FROM dual
UNION ALL SELECT '2016-01-02 04', '홍길동', '글등록'  , 1000 FROM dual
UNION ALL SELECT '2016-01-02 05', '홍길동', '댓글등록',  500 FROM dual
UNION ALL SELECT '2016-01-03 06', '홍길동', '글등록'  , 1000 FROM dual
UNION ALL SELECT '2016-01-05 07', '홍길동', '전환신청', 4000 FROM dual
UNION ALL SELECT '2016-01-07 08', '홍길동', '글등록'  , 1000 FROM dual
)
SELECT reg_dt, id, gb, milg
     , SUM(DECODE(gb, '전환신청',     0, milg))
       OVER(PARTITION BY id ORDER BY reg_dt) milg_tot
     , SUM(DECODE(gb, '전환신청', -milg, milg))
       OVER(PARTITION BY id ORDER BY reg_dt) milg_rem
  FROM t
;

마농님 감사합니다.

많은 도움이 되었습니다.

sum over를 쓰면 오히려 더 쉽게 구해졌네요..ㅠㅠ