case 문 도배가 성능상 제일 유리할 것으로 보입니다. 

더 좋은 방법도 많겠지만

CASE문 아닌 ROW_NUMBER()를 사용해서 한번 해봤습니다.

;
WITH TBL1 AS
(
     SELECT 10 AS C1, 11 AS C2, 9 AS C3 UNION ALL
     SELECT 10 AS C1, 11 AS C2, 9 AS C3 UNION ALL
     SELECT 1 AS C1, 2 AS C2, 3 AS C3 UNION ALL
     SELECT 3 AS C1, 1 AS C2, 2 AS C3
)
, TBL2 AS
(
     SELECT ROW_NUMBER() OVER (ORDER BY C1) AS ROW_NUM
      , C1, C2, C3
     FROM TBL1
)

SELECT MAX(CASE WHEN ROW_NUM2 = 1 THEN C1 ELSE NULL END) AS C1
     , MAX(CASE WHEN ROW_NUM2 = 2 THEN C1 ELSE NULL END) AS C2
     , MAX(CASE WHEN ROW_NUM2 = 3 THEN C1 ELSE NULL END) AS C3  
FROM
(
     SELECT ROW_NUM
          , C1
          , ROW_NUMBER() OVER (PARTITION BY ROW_NUM ORDER BY C1) AS ROW_NUM2
     FROM
     (
          SELECT ROW_NUM, C1 FROM TBL2
          UNION ALL
          SELECT ROW_NUM, C2 FROM TBL2
          UNION ALL
          SELECT ROW_NUM, C3 FROM TBL2
     ) AS B
) AS A
GROUP BY ROW_NUM