SELECT MAX(CASE WHEN Rn=1 THEN DateStamp END) AS DateStamp1,
MAX(CASE WHEN Rn=1 THEN Reading END) AS Reading1,
MAX(CASE WHEN Rn=1 THEN Channel END) AS Channel1,
MAX(CASE WHEN Rn=2 THEN DateStamp END) AS DateStamp2,
MAX(CASE WHEN Rn=2 THEN Reading END) AS Reading2,
MAX(CASE WHEN Rn=2 THEN Channel END) AS Channel2
FROM
(
SELECT ROW_NUMBER() OVER (PARTITION BY Grp ORDER BY Channel) AS Rn,*
FROM
(
SELECT DENSE_RANK() OVER(PARTITION BY Channel ORDER BY Reading) AS Grp,*
FROM table
)t
)r
GROUP BY Grp

with yourTable (DateStamp , Reading , Channel) as (
select '01-06-2012' , 1.0 , 111 union all
select '01-06-2012' , 2.0 , 111 union all
select '01-06-2012' , 3.0 , 222 union all
select '01-06-2012' , 4.0 , 222 )


select dateStamp
      ,[111]
      ,[222]
from  (
      select DateStamp
            ,Reading
            ,Channel
            ,row_number() over (partition by channel order by reading) rn
      from  yourTable
      ) d
pivot (max(Reading) for channel in ([111],[222])) p

OUTPUT:
dateStamp  111                                     222
---------- --------------------------------------- ---------------------------------------
01-06-2012 1.0                                     3.0
01-06-2012 2.0                                     4.0

quote:

---

Originally posted by ab2121

For my case the database has a very large number of rows but still only 3 columns
would it be easier (or possible) to format it like this :

Datestamp | Reading (from channel 111) | Reading (from Channel 121)| Reading (from Channel 131)|... etc

The maximum number of channels chosen would be 5

---