기능을 제한하면 쿼리만 복잡해집니다.
Sql 사용 기능에 제한을 두는 이유가 뭔가요?
낮은 버전 때문인가요?
낮은 버전때문이라면 기능의 한계로 어쩔 수 없다지만
그게 아닌 다른 이유라면? 어디까지 제한을 둬야 하는지..
말씀하신 것 빼고는 다써도 되는건지요?
순수 쿼리의 범주가 어디까지인지?
case문이나 rollup 등을 사용해도 되나요?

제가 테스트를 해본결과...

case문과 rollup()문은 허용이 되었습니다..

그러나, select ~ from 사이에 서브쿼리삽입은 허용이 안되었구요..ㅠㅠ

rank() over() 도 안되더군요...ㅠㅠ(당연히 partition by도 안먹힙니다....)

WITH members AS
(
SELECT '201004' dt, 100 cnt FROM dual
UNION ALL SELECT '201005',  50 FROM dual
UNION ALL SELECT '201006', 130 FROM dual
UNION ALL SELECT '201007', 140 FROM dual
UNION ALL SELECT '201008',  55 FROM dual
)
SELECT DECODE(rn, 1, '신규회원수'
                , 2, '전월대비 증감율'
                , 3, '누적 회원수'    ) AS gubun
     , MIN(DECODE(dt, '201004', v)) "201004"
     , MIN(DECODE(dt, '201005', v)) "201005"
     , MIN(DECODE(dt, '201006', v)) "201006"
     , MIN(DECODE(dt, '201007', v)) "201007"
     , MIN(DECODE(dt, '201008', v)) "201008"
  FROM (SELECT a.dt
             , b.rn
             , DECODE(b.rn, 1, cnt, 2, rate, 3, scnt) AS v
          FROM (SELECT a.dt
                     , a.cnt cnt
                     , ROUND(a.cnt / SUM(b.cnt), 5) rate
                     , a.cnt + NVL(SUM(b.cnt), 0) scnt
                  FROM members a
                     , members b
                 WHERE a.dt > b.dt(+)
                 GROUP BY a.dt, a.cnt
                 ORDER BY a.dt
                ) a
             , (SELECT ROWNUM rn FROM dict WHERE ROWNUM <= 3) b
        )
 GROUP BY rn
 ORDER BY rn
;

WITH members AS
(
SELECT '201004' dt, 100 cnt FROM dual
UNION ALL SELECT '201005',  50 FROM dual
UNION ALL SELECT '201006', 130 FROM dual
UNION ALL SELECT '201007', 140 FROM dual
UNION ALL SELECT '201008',  55 FROM dual
)
SELECT DECODE(gid, 0, '신규회원수'
                 , 1, '전월대비 증감율'
                 , 3, '누적 회원수'    ) AS gubun
     , DECODE(gid, 0, c1, 1, ROUND(c1/s1,5), 3, c1+NVL(s1,0)) "201004"
     , DECODE(gid, 0, c2, 1, ROUND(c2/s2,5), 3, c2+NVL(s2,0)) "201005"
     , DECODE(gid, 0, c3, 1, ROUND(c3/s3,5), 3, c3+NVL(s3,0)) "201006"
     , DECODE(gid, 0, c4, 1, ROUND(c4/s4,5), 3, c4+NVL(s4,0)) "201007"
     , DECODE(gid, 0, c5, 1, ROUND(c5/s5,5), 3, c5+NVL(s5,0)) "201008"
  FROM (SELECT GROUPING_ID(1, 2) gid
             , MIN(DECODE(dt, '201004', cnt)) c1
             , MIN(DECODE(dt, '201005', cnt)) c2
             , MIN(DECODE(dt, '201006', cnt)) c3
             , MIN(DECODE(dt, '201007', cnt)) c4
             , MIN(DECODE(dt, '201008', cnt)) c5
             , SUM(CASE WHEN dt < '201004' THEN cnt END) s1
             , SUM(CASE WHEN dt < '201005' THEN cnt END) s2
             , SUM(CASE WHEN dt < '201006' THEN cnt END) s3
             , SUM(CASE WHEN dt < '201007' THEN cnt END) s4
             , SUM(CASE WHEN dt < '201008' THEN cnt END) s5
          FROM members
         GROUP BY ROLLUP(1, 2)
        )
;