Reputation: 3303
This is my data. I'm trying to convert vertical data to horizontal data. I tried using PIVOT, but I don't have a column to be used as aggregate.
I don't need to use PIVOT, but it's the operator I've used for this kind of thing. I also considered it since I know all possible GroupId
s.
Name GroupId
Joe B1
John B2
Mary C1
Lisa D2
Joe D2
The result would be something like this. I have to concatenate all the strings for each name:
Name B1 B2 C1 D2
Joe B1 D2
John B2
Mary C1
Lisa D2
I was thinking of something like this, but I knew it will not produce the desired results.
declare @table table
(
Name varchar(10),
GroupId varchar(2)
)
insert into @table
select 'Joe', 'B1' union
select 'John','B2' union
select 'Mary','C1' union
select 'Lisa','D2' union
select 'Joe','D2'
select *
from
(
select Name, GroupId
from @table
) src
pivot
(
min(Name)
for GroupId in ([B1], [B2], [C1], [D2])
) piv;
Upvotes: 3
Views: 468
Reputation: 1955
;WITH CTE AS (SELECT Name, GroupId, 1 id FROM @table)
select Name
, [B1] = CASE WHEN [B1] IS NULL THEN '' ELSE 'B1' END
, [B2] = CASE WHEN [B2] IS NULL THEN '' ELSE 'B2' END
, [C1] = CASE WHEN [C1] IS NULL THEN '' ELSE 'C1' END
, [D2] = CASE WHEN [D2] IS NULL THEN '' ELSE 'D2' END
from CTE
PIVOT (min(ID) for GroupId in ([B1], [B2], [C1], [D2])) piv;
OUTPUT
Name B1 B2 C1 D2
---------- ---- ---- ---- ----
Joe B1 D2
John B2
Lisa D2
Mary C1
(4 rows affected)
Upvotes: 0
Reputation: 50163
You can use conditional aggregation instead
select Name,
max(case when GroupId = 'B1' then GroupId end) [B1],
max(case when GroupId = 'B2' then GroupId end) [B2],
max(case when GroupId = 'C1' then GroupId end) [C1],
max(case when GroupId = 'D2' then GroupId end) [D2]
from @table t
group by Name;
For your current attempt with pivot
operator do aggregation with only one column which is GroupId
select *
from
(
select Name, GroupId
from @table
) src pivot (
max(GroupId)
for GroupId in ([B1], [B2], [C1], [D2])
) piv;
Upvotes: 3