T-SQL Query to convert rows to coumns based on mutiple tables - tsql

Question

I have two master tables CompanyMaster, ActivityMaster for a child table CompanyActivities
ActivityMaster
ACTIVITYID ACTIVITYNAME
A1 testActivity
A2 someActivity
A3 otheractivity
A4 someotheractivity
A5 anyotheractivity
CompanyMaster
COMPANYID COMPANYNAME
C1 testcompany
C2 ACompany
C3 MyCompany
C4 SomeCompany
C5 ZCompany
C6 Company123
C7 ComapnyABC
CompanyActivities - The COMPANYID in CompanyActivities is having a primarykey-foreighkey relation ship with COMPANYID in CompanyMaster (primary key table) and ACTIVITYID is having a primarykey-foreighkey relation ship with ACTIVITYID in ActivityMaster(primary key table)
COMPANYID ACTIVITYID
C1 A1
C1 A3
C3 A1
C3 A2
C4 A5
C5 A1
C6 A3
C7 A3
I want to do write a query to get the following output where all the rows in ACTIVITYID column of the ActivityMaster table will be converted to columns
Output
Companies A1 A2 A3 A4 A5
C1 Y N Y N N
C2 N N N N N
C3 Y Y N N N
C4 N N N N Y
C5 Y N N N N
C6 N N Y N N
C7 N N Y N N
The output table displays all the companies as rows in the first column and all the activities are shown as columns that start after the first column, if there is row that contains both ACTIVITYID and COMPANYID it will set to Y in output otherwise it would be set to N
eg- COMPANYID C1 is having an activity ACTIVITYID A1 in CompanyActivities table so the first row in the second column that comes just below A1 and in the right to C1 is set Y, whereas C1 and A2 are not having a row, so the third column in the first row is set to N
I am using C#.net and 4 for loops to achieve the output now which is talking a heavy toll on the performance of the application, So i would like to do this using a query, I have searched for pivot queries, but all the examples i found knows the column names before-hand, which i don't i only get the names of the column names by querying the ActivityMaster.

Answer 1

create table #CompanyMaster (COMPANYID int, COMPANYNAME varchar(30))
create table #ActivityMaster (ACTIVITYID int, ACTIVITYNAME varchar(30))
create table #CompanyActivities (COMPANYID int, ACTIVITYID int)
insert into #CompanyMaster
SELECT 1, 'Company A'
union all
SELECT 2, 'Company B'
insert into #ActivityMaster
SELECT 101, 'Activity X'
union all
SELECT 102, 'Activity Y'
union all
SELECT 103, 'Activity Z'
insert into #CompanyActivities
select 1, 102
union all
select 2, 101
-- build activities column names
--case [Activity X] when 0 then ''N'' else ''Y'' end as [Activity X],
--case [Activity Y] when 0 then ''N'' else ''Y'' end as [Activity Y],
--case [Activity Z] when 0 then ''N'' else ''Y'' end as [Activity Z]
declare #activities nvarchar(max)
set #activities
= (
select 'case [' + ACTIVITYNAME + '] when 0 then ''N'' else ''Y'' end as [' + ACTIVITYNAME + '],' + char(10)
from #ActivityMaster
for xml path('')
)
set #activities = substring(#activities, 0, len(#activities)-1)
declare #activities_for nvarchar(max)
-- build activities column names in for
--[Activity X], [Activity Y], [Activity Z]
set #activities_for
= (
select '[' + ACTIVITYNAME + '],' + char(10)
from #ActivityMaster
for xml path('')
)
set #activities_for = substring(#activities_for, 0, len(#activities_for)-1)
declare #sql nvarchar(MAX) = N'
select COMPANYNAME,
<activities>
From
(select c.COMPANYNAME, a.ACTIVITYNAME,
(case
when ca.ACTIVITYID is not null and ca.COMPANYID is not null then 1
else 0
end) as STATUS
from #CompanyMaster c
cross join #ActivityMaster a
left join #CompanyActivities ca on ca.COMPANYID = c.COMPANYID and a.ACTIVITYID = ca.ACTIVITYID) p
pivot
(
sum(STATUS) for ACTIVITYNAME IN (<activities_for>)
) as pvt
'
set #sql = replace(#sql, '<activities>', #activities)
set #sql = replace(#sql, '<activities_for>', #activities_for)
print #sql
exec sp_executesql #sql
drop table #CompanyMaster
drop table #ActivityMaster
drop table #CompanyActivities