얼마 전에 비슷하게 작성한 쿼리가 있어서 참고해서 작성했습니다. 
참고만 하시구요.. ㅋㅋ
더 효율적인 쿼리는 다음분이 올려주실 거에요.. ^^;;  

WITH TMP AS (

        SELECT '홍길동' NAME, '20121122083105' DATE, '15000' PRICE FROM DUAL 

    UNION ALL SELECT '홍길동' NAME, '20121122083115' DATE, '10000' PRICE FROM DUAL 

    UNION ALL SELECT '고길동' NAME, '20121122083211' DATE, '12000' PRICE FROM DUAL 

    UNION ALL SELECT '홍길동' NAME, '20121122084550' DATE, '11000' PRICE FROM DUAL 

    UNION ALL SELECT '고길동' NAME, '20121122091103' DATE, '19000' PRICE FROM DUAL 

    UNION ALL SELECT '고길동' NAME, '20121122103918' DATE, '21000' PRICE FROM DUAL 

)

SELECT 

    NAME, 

    RN - NVL(LAG(RN) OVER(ORDER BY DATE DESC), 0) CNT

FROM (

    SELECT  

        NAME, DATE, PRICE, 

        DECODE(

            LEAD(NAME) OVER(ORDER BY DATE DESC), 

            NAME, 0, 

            ROW_NUMBER() OVER(ORDER BY DATE DESC)

        ) RN

    FROM TMP 

    ORDER BY DATE DESC 

)

WHERE RN > 0 

SELECT name
     , COUNT(*) cnt
--     , MIN(dt) dt_from
--     , MAX(dt) dt_to
--     , SUM(price) price
  FROM (SELECT name, dt, price
             , ROW_NUMBER() OVER(ORDER BY dt DESC) rn1
             , ROW_NUMBER() OVER(PARTITION BY name ORDER BY dt DESC) rn2
          FROM t
        )
 GROUP BY name, rn1-rn2
 ORDER BY MAX(dt) DESC
;

많은 도움이 되었습니다.  감사합니다