Converting multiple rows to columns with specified heading - sql-server

Question

I have a table that holds details of activities carried out by individuals - contents of this table is similar to the following:
| Person | Category | Activity |
--------------------------------------
| Username1 | X | X1 |
| Username1 | Y | Y1 |
| Username1 | Z | Z1 |
I need a SQL query that can produce something like the following and any help would be appreciated:
| Person | Cat1 | Cat1_Act|Cat2 | Cat2_Act| Cat3 | Cat3_Act |
---------------------------------------------------------------
| Username1 | X | X1 | Y | Y1 | Z | Z1 |
I understand reading through a number of posts that PIVOT can be used to achieve this but I have not been to find a solution close to what I need as most solutions are often to use values e.g 'X', 'Y', 'Z' (in my example table) as table headers but I want to ideally be able to specify name for the table headers holding the new columns (Hope this all makes sense and someone can help :-) )

Answer 1

There are several ways that you can get the desired result. If you have a limited number of values that you want to PIVOT into columns, then you can hard-code the query a few different ways.
Aggregate function with CASE:
select
person,
max(case when seq = 1 then category end) Cat1,
max(case when seq = 1 then activity end) Cat1_Act,
max(case when seq = 2 then category end) Cat2,
max(case when seq = 2 then activity end) Cat2_Act,
max(case when seq = 3 then category end) Cat3,
max(case when seq = 3 then activity end) Cat3_Act
from
(
select person, category, activity,
row_number() over(partition by person
order by category) seq
from yourtable
) d
group by person;
See SQL Fiddle with Demo. By assigning a sequence or row_number to each category per user, you can use this row number to convert the rows into columns.
Static PIVOT:
If you want to apply the PIVOT function, then I would first suggest unpivoting the category and activity columns into multiple rows and then apply the pivot function.
;with cte as
(
select person, category, activity,
row_number() over(partition by person
order by category) seq
from yourtable
)
select person,
cat1, cat1_act,
cat2, cat2_act,
cat3, cat3_act
from
(
select t.person,
col = case
when c.col = 'cat' then col+cast(seq as varchar(10))
else 'cat'+cast(seq as varchar(10))+'_'+col
end,
value
from cte t
cross apply
(
select 'cat', category union all
select 'act', activity
) c (col, value)
) d
pivot
(
max(value)
for col in (cat1, cat1_act, cat2, cat2_act,
cat3, cat3_act)
) piv;
See SQL Fiddle with Demo
Dynamic PIVOT: Finally if you have an unknown number of values then you can use dynamic SQL to get the result:
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX)
select #cols = STUFF((SELECT ','
+ QUOTENAME(case
when d.col = 'cat' then col+cast(seq as varchar(10))
else 'cat'+cast(seq as varchar(10))+'_'+col end)
from
(
select row_number() over(partition by person
order by category) seq
from yourtable
) t
cross apply
(
select 'cat', 1 union all
select 'act', 2
) d (col, so)
group by col, so, seq
order by seq, so
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query = 'SELECT person, ' + #cols + '
from
(
select t.person,
col = case
when c.col = ''cat'' then col+cast(seq as varchar(10))
else ''cat''+cast(seq as varchar(10))+''_''+col
end,
value
from
(
select person, category, activity,
row_number() over(partition by person
order by category) seq
from yourtable
) t
cross apply
(
select ''cat'', category union all
select ''act'', activity
) c (col, value)
) x
pivot
(
max(value)
for col in (' + #cols + ')
) p '
execute sp_executesql #query;
See SQL Fiddle with Demo. All versions give a result:
| PERSON | CAT1 | CAT1_ACT | CAT2 | CAT2_ACT | CAT3 | CAT3_ACT |
| Username1 | X | X1 | Y | Y1 | Z | Z1 |

Answer 2

this is a simple example
SELECT
Person,
MAX(CASE Category WHEN 'X' THEN Activity ELSE 0 END) AS 'X'
MAX(CASE Category WHEN 'Y' THEN Activity ELSE 0 END) AS 'Y'
MAX(CASE Category WHEN 'Z' THEN Activity ELSE 0 END) AS 'Z'
FROM mytable
GROUP BY Person