create table testf1(f1 char(1))
insert into testf1 
select 'a' union all
select 'b'union all
select 'c'union all
select 'd'union all
select 'e'union all
select 'f'union all
select 'g'union all
select 'h'union all
select 'i'union all
select 'j';

--sql 2005
;with cte as
(
    select f1,rid=row_number() over (order by getdate())
    from testf1
)

select a.f1 as af,b.f1 as bf,c.f1 as cf
from (select *,(rid-1)/3 as fg from cte where (rid-1)%3=0) a
    left join (select *,(rid-1)/3 as fg from cte where (rid-1)%3=1) b on a.fg = b.fg
    left join (select *,(rid-1)/3 as fg from cte where (rid-1)%3=2) c on a.fg = c.fg

--sql 2000

select f1,rid=identity(int,1,1) into #testf1 from testf1

select a.f1 as af,b.f1 as bf,c.f1 as cf
from (select *,(rid-1)/3 as fg from #testf1 where (rid-1)%3=0) a
    left join (select *,(rid-1)/3 as fg from #testf1 where (rid-1)%3=1) b on a.fg = b.fg
    left join (select *,(rid-1)/3 as fg from #testf1 where (rid-1)%3=2) c on a.fg = c.fg

drop table #testf1
drop table testf1

/*****************************

(10 行受影响)
af   bf   cf
---- ---- ----
a    b    c
d    e    f
g    h    i
j    NULL NULL

(4 行受影响)

(10 行受影响)

af   bf   cf
---- ---- ----
a    b    c
d    e    f
g    h    i
j    NULL NULL

(4 行受影响)

------解决方案--------------------
SQL code

DECLARE @A TABLE([F1] VARCHAR(1))
INSERT @A
SELECT 'A' UNION ALL
SELECT 'B' UNION ALL
SELECT 'C' UNION ALL
SELECT 'D' UNION ALL
SELECT 'E' UNION ALL
SELECT 'F' UNION ALL
SELECT 'G' UNION ALL
SELECT 'H' UNION ALL
SELECT 'I' UNION ALL
SELECT 'J'

;WITH M1 AS
(
    SELECT ROW_NUMBER() OVER ( ORDER BY (SELECT 1)) AS ID,* FROM @A
),M2 AS
(
    SELECT 
        ID%3 AS RID,
        ROW_NUMBER () OVER (PARTITION BY ID%3 ORDER BY ID) AS NID,
    * FROM M1
)
SELECT 
    MAX(CASE WHEN RID=1 THEN F1 ELSE '' END) AS F1,
    MAX(CASE WHEN RID=2 THEN F1 ELSE '' END) AS F2,
    MAX(CASE WHEN RID=0 THEN F1 ELSE '' END) AS F3
FROM M2 GROUP BY NID
/*
F1   F2   F3
---- ---- ----
A    B    C
D    E    F
G    H    I
J         
*/