오라클은 안 써봤지만.. 이런 식으로 해볼 수 있을 듯 하네요.

WITH T AS (
    SELECT 'DATE' AS DT , '2000-01-01' COL1, NULL COL2, NULL COL3 FROM DUAL UNION ALL
    SELECT 'DATE' AS DT , '2000-02-02' COL1, NULL COL2, NULL COL3 FROM DUAL UNION ALL
    SELECT 'DATE' AS DT ,  NULL COL1, '2019-02-05' COL2, NULL COL3 FROM DUAL UNION ALL
    SELECT 'DATE' AS DT ,  NULL COL1, '2019-02-05' COL2, NULL COL3 FROM DUAL UNION ALL
    SELECT 'DATE' AS DT ,  NULL COL1, '2019-02-12' COL2, NULL COL3 FROM DUAL UNION ALL
    SELECT 'DATE' AS DT ,  NULL COL1, '2019-02-20' COL2, NULL COL3 FROM DUAL UNION ALL
    SELECT 'DATE' AS DT ,  NULL COL1, NULL COL2, '2019-02-05' COL3 FROM DUAL UNION ALL
    SELECT 'DATE' AS DT ,  NULL COL1, NULL COL2, '3030-01-01' COL3 FROM DUAL UNION ALL
    SELECT 'DATE' AS DT ,  NULL COL1, NULL COL2, '2019-02-28' COL3 FROM DUAL UNION ALL
    SELECT 'DATE' AS DT ,  NULL COL1, NULL COL2, '2019-02-28' COL3 FROM DUAL UNION ALL
    SELECT 'DATE2' AS DT , '2018-09-01' COL1, NULL COL2, NULL COL3 FROM DUAL UNION ALL
    SELECT 'DATE2' AS DT , '2019-02-01' COL1, NULL COL2, NULL COL3 FROM DUAL UNION ALL
    SELECT 'DATE2' AS DT ,  NULL COL1, '2019-02-07' COL2, NULL COL3 FROM DUAL UNION ALL
    SELECT 'DATE2' AS DT ,  NULL COL1, NULL COL2, '2016-02-15' COL3 FROM DUAL UNION ALL
    SELECT 'DATE2' AS DT ,  NULL COL1, NULL COL2, '2016-03-01' COL3 FROM DUAL UNION ALL
    SELECT 'DATE2' AS DT ,  NULL COL1, NULL COL2, '2016-03-15' COL3 FROM DUAL UNION ALL
    SELECT 'DATE2' AS DT ,  NULL COL1, NULL COL2, '2016-03-28' COL3 FROM DUAL
), C1 (SEQ, DT, COL1) AS (
    SELECT ROW_NUMBER() OVER(PARTITION BY DT ORDER BY COL1), DT, COL1 FROM T WHERE COL1 IS NOT NULL
), C2 (SEQ, DT, COL2) AS (
    SELECT ROW_NUMBER() OVER(PARTITION BY DT ORDER BY COL2), DT, COL2 FROM T WHERE COL2 IS NOT NULL
), C3 (SEQ,DT,  COL3) AS
    SELECT ROW_NUMBER() OVER(PARTITION BY DT ORDER BY COL3), DT, COL3 FROM T WHERE COL3 IS NOT NULL
)
SELECT COALESCE(C1.DT, C2.DT, C3.DT) AS DT, C1.COL1, C2.COL2, C3.COL3
FROM C1
    FULL OUTER JOIN C2 ON C2.DT = C1.DT AND C2.SEQ = C1.SEQ
    FULL OUTER JOIN C3 ON C3.DT = C2.DT AND C3.SEQ = C2.SEQ
ORDER BY COALESCE(C1.DT, C2.DT, C3.DT)
    , COALESCE(C1.SEQ, C2.SEQ, C3.SEQ);

select *
from (select dt, dts
           , gb
           , row_number() over(partition by dt, gb order by rn) pgb
      from (select dt
                 , col1||col2||col3 as dts
                 , case when col1 is not null then 1
                        when col2 is not null then 2
                        when col3 is not null then 3
                   end gb
                 , rownum rn
            from t
           )
      )
pivot(min(dts) as dtss for(gb) in (1, 2, 3))
order by dt, pgb
;