Need Helping hand on below result

Question

Hi All,
Need a  helping hand to get below result in Sql server query.
Source table data:
SW
SM
SaName
Commitment
ID
QP
Week 2 October, 2014
October, 2014
Solution
Uncommitted Upside
6-53I6YHDDI
311219.6888
Week 3 October, 2014
October, 2014
Solution
Uncommitted Upside
6-53I6YHDDI
483853.9999
Week 4 October, 2014
October, 2014
Solution
Uncommitted Upside
6-53I6YHDDI
483853.9999
Week 5 October, 2014
October, 2014
Solution
Uncommitted Upside
6-53I6YHDDI
483853.9999
Week 1 November, 2014
November, 2014
Proof
Uncommitted Upside
6-53I6YHDDI
483853.9999
Expecting Result in a single record for ID
Result
Week 2 October, 2014 QP
Week2 CommitmentLevel
Week 2 October, 2014 SalesStageName
Week 3 October, 2014 QP
Week3 CommitmentLevel
Week 3 October, 2014 SalesStageName
Week 4 October, 2014 QP
Week4 CommitmentLevel
Week 4 October, 2014 SalesStageName
6-53I6YHDDI
311219.6888
Uncommitted Upside
Solution
483853.9999
Uncommitted Upside
Solution
483853.9999
Uncommitted Upside
Solution
Thanks in Advance...

Answer 1

vgvchinna,
See below two ways to get data in your desired format.
The 1st is a simple OUTER JOIN after partitioning the data using CTE.
The 2nd introduces the SW date column three times with aliases to use with three PIVOTs for each column of data. It then wraps the results in a MAX+GROUP BY to flatten it out.
I can think of a 3rd approach using CASE statements too but not including the SQL code for that here.
As you can see both these approaches are complex and not easily scalable. There is also dependency on you knowing the values beforehand to put in the WHERE clause (1st approach) or the PIVOT (in 2nd approach). Removing this dependency will
require dynamic SQL which will complicate the solution even more.
Let us know in this forum if you'd like any explanation on these two approaches.
On a side note, I'd also suggest that when you ask a question in any forum, include as much info as possible e.g. the CREATE/INSERT scripts, properly formatted desired output, what approaches have you already tried, your comfort
and skill level with the technology etc. so that you get better answers and also save time for the people who respond to you.
Good luck!
declare @t table
ID varchar(20)
, SW varchar(25)
, QP varchar(20)
, CommitmentLevel varchar(20)
, SaleStageName varchar(20)
insert into @t
values
('6-53I6YHDDI', 'Week 2 October, 2014', '311219.6888', 'Uncommitted Upside', 'Solution'),
('6-53I6YHDDI', 'Week 3 October, 2014', '483853.9999', 'Uncommitted Upside', 'Solution'),
('6-53I6YHDDI', 'Week 4 October, 2014', '483853.9999', 'Uncommitted Upside', 'Solution'),
('6-53I6YHDDI', 'Week 1 November, 2014', '483853.9999', 'Uncommitted Upside', 'Proof')
select * from @t--------------------
-- Method 1 with CTE
;with
wk2 as (select ID, QP as wk2QP, CommitmentLevel as wk2CommitmentLevel, SaleStageName as wk2SaleStageName from @t where SW = 'Week 2 October, 2014'),
wk3 as (select ID, QP as wk3QP, CommitmentLevel as wk3CommitmentLevel, SaleStageName as wk3SaleStageName from @t where SW = 'Week 3 October, 2014'),
wk4 as (select ID, QP as wk4QP, CommitmentLevel as wk4CommitmentLevel, SaleStageName as wk4SaleStageName from @t where SW = 'Week 3 October, 2014'),
wk1Nov as (select ID, QP as wk1NovQP, CommitmentLevel as wk1NovCommitmentLevel, SaleStageName as wk1NovSaleStageName from @t where SW = 'Week 1 November, 2014')
select
coalesce(wk2.ID, wk3.ID, wk4.ID, wk1Nov.ID) as Result,
wk2QP, wk2CommitmentLevel, wk2SaleStageName,
wk3QP, wk3CommitmentLevel, wk3SaleStageName,
wk4QP, wk4CommitmentLevel, wk4SaleStageName,
wk1NovQP, wk1NovCommitmentLevel, wk1NovSaleStageName
from wk2 full outer join wk3
on wk2.ID = wk3.ID
full outer join wk4
on wk2.ID = wk4.ID
full outer join wk1Nov
on wk2.ID = wk1Nov.ID
-- Method 2 with PIVOT
select ID
, max([Week 2 October, 2014QP]) as wk2QP, max([Week 2 October, 2014CommitmentLevel]) as wk2CommitmentLevel, max([Week 2 October, 2014SaleStageName]) as wk2SaleStageName
, max([Week 3 October, 2014QP]) as wk3QP, max([Week 3 October, 2014CommitmentLevel]) as wk3CommitmentLevel, max([Week 3 October, 2014CommitmentLevel]) as wk3SaleStageName
, max([Week 4 October, 2014QP]) as wk4QP, max([Week 4 October, 2014CommitmentLevel]) as wk4CommitmentLevel, max([Week 4 October, 2014QP]) as wk3SaleStageName
, max([Week 1 November, 2014QP]) as wk1NovQP, max([Week 1 November, 2014CommitmentLevel]) as wk1NovCommitmentLevel, max([Week 1 November, 2014QP]) as wk1NovSaleStageName
from
( select ID, SW+'QP' as SWQP, QP, SW+'CommitmentLevel' as SWCommitmentLevel, CommitmentLevel, SW+'SaleStageName' as SWSaleStageName, SaleStageName from @t ) as pivotquery
PIVOT ( MAX(QP) for SWQP in ([Week 2 October, 2014QP], [Week 3 October, 2014QP], [Week 4 October, 2014QP], [Week 1 November, 2014QP]) ) as pvtQP
PIVOT ( MAX(CommitmentLevel) for SWCommitmentLevel in ([Week 2 October, 2014CommitmentLevel], [Week 3 October, 2014CommitmentLevel], [Week 4 October, 2014CommitmentLevel], [Week 1 November, 2014CommitmentLevel]) ) as pvtCommitmentLevel
PIVOT ( MAX(SaleStageName) for SWSaleStageName in ([Week 2 October, 2014SaleStageName], [Week 3 October, 2014SaleStageName], [Week 4 October, 2014SaleStageName], [Week 1 November, 2014SaleStageName]) ) as pvtSaleStage
group by ID
- Aalamjeet Rangi | (Blog)