Merge two or more columns dynamically based on table columns? - sql

Question

I have a table which is called SectionNames as follows
SectionID SectionCode Subsection
1 xYz Individual
2 xYz Family
3 CYD Friends
4 PCPO level1
5 PCPO level2
6 PCPO level3
So on. So in future we can add one or more subsections for each section code.
And have one more table which is a reference table for above SectionNames table and Employee table with employee data.
ID EmployeeID SectionID Cost
1 1 1 $200
2 1 2 $300
3 1 3 $40
4 1 4 $10
5 1 5 No Level
6 1 6 No Level
7 1 7 $20
8 1 8 No Level
9 1 9 No Level
So Iwant the out put from these tables should look like:
EmployeeID Individual_xyz_Cost Family_xyz_Cost Friends_xyz_cost level1_PCPO_cost level2_PCPO_Cost
1 $200 $300 $400 $10 NoLevel
There are few employee records exists in my employee table. And I want this to be dynamic. Like If in future if one more subsection called Relatives added for XYZ section then my query should return Relatives_XYZ_Cost.
How can I write this query dynamically?

Answer 1

You will want to use the PIVOT function to transform the data from columns into rows. If you are going to have an unknown number of values that need to be columns, then you will need to use dynamic SQL.
It is easier to see a static or hard-coded version first and then convert it into a dynamic SQL version. A static version is used when you have a known number of values:
select *
from
(
select e.employeeid,
s.subsection +'_'+s.sectioncode+'_Cost' Section,
e.cost
from employee e
inner join sectionnames s
on e.sectionid = s.sectionid
) src
pivot
(
max(cost)
for section in (Individual_xYz_Cost, Family_xYz_Cost,
Friends_CYD_Cost, level1_PCPO_Cost,
level2_PCPO_Cost, level3_PCPO_Cost)
) piv;
See SQL Fiddle with Demo.
If you need the query to be flexible, then you will convert this to use dynamic SQL:
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX)
select #cols = STUFF((SELECT ',' + QUOTENAME(subsection +'_'+sectioncode+'_Cost')
from SectionNames
group by subsection, sectioncode, sectionid
order by sectionid
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query = 'SELECT employeeid,' + #cols + '
from
(
select e.employeeid,
s.subsection +''_''+s.sectioncode+''_Cost'' Section,
e.cost
from employee e
inner join sectionnames s
on e.sectionid = s.sectionid
) x
pivot
(
max(cost)
for section in (' + #cols + ')
) p '
execute(#query)
See SQL Fiddle with Demo
The result of both is:
| EMPLOYEEID | INDIVIDUAL_XYZ_COST | FAMILY_XYZ_COST | FRIENDS_CYD_COST | LEVEL1_PCPO_COST | LEVEL2_PCPO_COST | LEVEL3_PCPO_COST |
----------------------------------------------------------------------------------------------------------------------------------
| 1 | $200 | $300 | $40 | $10 | No Level | No Level |