SQL Server Sort fixed number of rows/columns horizontally - sql

Question

I'm trying use PIVOT in a SQL Server stored procedure to take the following data:
ID Rank
203081 1.1
200761 3.9
202687 5.3
203135 5.0
203090 3.3
and return the ID's sorted horizontally. The ranking goes from 1 to 6 with each rank having tenths in between. Example 1.0, 1.1, 1.2, 1.3, ... 1.9, 2.0
The ID's need to be sorted in order by rank.
The result should return something like the following:
(if additional columns needed for aggregation that is fine as well.)
[1] ,[2] ,[3] ,[4] ,[5]
202687,203135,200761,203090,203081
Using above data the ID's would be sorted by rank as 5.3->5.0->3.9->3.3->1.1
In the end I need to take the results and insert them into another table with the ID's sorted horizontally.
I can't get the PIVOT to work correctly. I'm sure it is something obvious I'm not seeing.
If there is a better/faster way to achieve what is needed I would like to know what that solution would be as well.

Answer 1

It sounds like you just need to use row_number() to get the ordering correct and then PIVOT the data on that row number.
If you have a limited number values then you can use:
select [1], [2], [3], [4], [5]
from
(
select id,
row_number() over(order by [rank] desc) seq
from yourtable
) d
pivot
(
max(id)
for seq in ([1], [2], [3], [4], [5])
) piv;
See SQL Fiddle with Demo
Or if you are going to have an unknown number, you will use dynamic SQL:
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX)
select #cols = STUFF((SELECT distinct ',' + QUOTENAME(row_number() over(order by [rank] desc))
from yourtable
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query = 'SELECT ' + #cols + '
from
(
select id,
row_number() over(order by [rank] desc) seq
from yourtable
) x
pivot
(
max(id)
for seq in (' + #cols + ')
) p '
execute sp_executesql #query;
See SQL Fiddle with Demo. These give a result:
| 1 | 2 | 3 | 4 | 5 |
|--------|--------|--------|--------|--------|
| 202687 | 203135 | 200761 | 203090 | 203081 |