如何取消透视表中的多个列

Question

I have a table with below structure, I need to unpivot the output so that I get one row per ID per PARAMETER and it's corresponding RATINGS

create table RATINGS(ID INT,PARAMETER1 INT,PARAMETER2 INT,PARAMETER3 INT,PARAMETER4 INT)
insert into RATINGS values(1000,1,3,2,1)
insert into RATINGS values(1002,2,3,3,2)
insert into RATINGS values(1007,3,3,2,1)
insert into RATINGS values(1015,1,3,1,3)
insert into RATINGS values(1019,3,2,1,1)

Expected output:

ID    PARAMETERS RATING
1000  PARAMETER1  1
1000  PARAMETER2  3
1000  PARAMETER3  2
1000  PARAMETER4  1
1002  PARAMETER1  2
1002  PARAMETER2  3
1002  PARAMETER3  3
1002  PARAMETER4  2
1007  PARAMETER1  3
1007  PARAMETER2  3
1007  PARAMETER3  2
1007  PARAMETER4  1
1015  PARAMETER1  1
1015  PARAMETER2  3
1015  PARAMETER3  1
1015  PARAMETER4  3
1019  PARAMETER1  3
1019  PARAMETER1  2
1019  PARAMETER1  1
1019  PARAMETER1  1

Later, I also need to filter so that I get only those rows that have ratings 1 and 2 . So the output then must be

ID    PARAMETERS RATING
1000  PARAMETER1  1
1000  PARAMETER3  2
1000  PARAMETER4  1
1002  PARAMETER1  2
1002  PARAMETER4  2
1007  PARAMETER3  2
1007  PARAMETER4  1
1015  PARAMETER1  1
1015  PARAMETER3  1
1019  PARAMETER1  2
1019  PARAMETER1  1
1019  PARAMETER1  1

I am able to get the first two ID and PARAMETERS column using below query:

select ID,[parameters] from RATINGS
unpivot
(
[value] for [PARAMETERS] in (PARAMETER1,PARAMETER2,PARAMETER3,PARAMETER4)
) unpvt

OUTPUT:

ID      PARAMETERS
1000    PARAMETER1
1000    PARAMETER2
1000    PARAMETER3
1000    PARAMETER4
1002    PARAMETER1
1002    PARAMETER2
1002    PARAMETER3
1002    PARAMETER4
1007    PARAMETER1
1007    PARAMETER2
1007    PARAMETER3
1007    PARAMETER4
1015    PARAMETER1
1015    PARAMETER2
1015    PARAMETER3
1015    PARAMETER4
1019    PARAMETER1
1019    PARAMETER2
1019    PARAMETER3
1019    PARAMETER4

Could someone please let me know how to get the Rating column?

Answer 1

Just add the value column:

select ID,[parameters], [rating] = value 
from RATINGS
unpivot
(
[value] for [PARAMETERS] in (PARAMETER1,PARAMETER2,PARAMETER3,PARAMETER4)
) unpvt
where value in (1,2);

Answer 2

You can unpivot with UNION

select id, 'PARAMETERS1' as parameters, parameter1 as rating 
from ratings
union
select id, 'PARAMETERS2' as parameters, parameter2 as rating 
from ratings
union
select id, 'PARAMETERS3' as parameters, parameter3 as rating 
from ratings
union
select id, 'PARAMETERS4' as parameters, parameter4 as rating 
from ratings
order by id, parameters

http://www.sqlfiddle.com/#!9/421f05/7

Then the second part just

with parmratings as (

    select id, 'PARAMETERS1' as parameters, parameter1 as rating 
    from ratings
    union
    select id, 'PARAMETERS2' . . . 
)
select * from parmratings where rating < 3

Answer 3

I would use CROSS APPLY :

SELECT r.id, rr.PARAMETERS, rr.RATING
FROM ratings r CROSS APPLY
     ( VALUES ([PARAMETERS1], 'PARAMETERS1'), . . .  ) rr(RATING, PARAMETERS)
ORDER BY r.id;