• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 712
  • Last Modified:

Pivot/Unpivot master detail?

Hello,

SQL Server 2008.

I have a master/detail relationship between two tables.  I need to write an export that returns the columns from the master table followed by the entries in the details table.

I guess i need to pivot/unpivot the details table?  The details table has two columns.  These are code and value.  However i dont know beforehand what the possible contents of code can be.

The output should look:

id,Name,Startdate,Code,Value,Code,Value,Code,Value,Code,Value,Code,Value,......

CREATE TABLE [dbo].[Master](
	[id] [int] NOT NULL,
	[Name] [varchar](50) NULL,
	[Startdate] [date] NULL)

CREATE TABLE [dbo].[Details](
	[id] [int] NOT NULL,
	[MasterId] [int] NULL,
	[Code] [nvarchar](20) NULL,
	[Value] [nvarchar](50) NULL)

Open in new window


Any ideas?
0
soozh
Asked:
soozh
1 Solution
 
David ToddSenior DBACommented:
Hi,

A pivot operator isn't going to give you Code Value Code Value etc ...

it could give you value1 as Code1, value2 as Code2, value3 as Code3

See this:
use ExpertsExchange
go

set ansi_nulls on
set quoted_identifier on
set ansi_padding on
go

if exists (
	select * from sys.objects 
	where 
		object_id = OBJECT_ID( N'dbo.master' ) and type in ( N'U' )
	)
	drop table dbo.master
go

create table dbo.master (
	MasterID int identity not null
	, name varchar( 50 ) not null
	, startdate datetime null
	)
go

if exists (
	select * from sys.objects 
	where 
		object_id = OBJECT_ID( N'dbo.detail' ) and type in ( N'U' )
	)
	drop table dbo.detail
go

create table dbo.detail (
	DetailID int identity not null
	, MasterID int not null
	, code varchar( 20 ) null
	, charvalue varchar( 50 ) null
	, intvalue int null
	)
go

insert dbo.Master( name, startdate )
	values( 'Fred', dateadd( day, datediff( day, 0, getdate()), 0 ))
	, ( 'Jimmy', dateadd( day, datediff( day, 0, getdate()), 0 - 1 ))
;

go

insert dbo.detail( MasterID, code, intvalue )
	select 
		MasterID
		, 'something'
		, abs( checksum( newid())) % 100 
	from dbo.master

	union all

	select 
		MasterID
		, 'something else'
		, abs( checksum( newid())) % 100 
	from dbo.master


	union all

	select 
		MasterID
		, 'another thing'
		, abs( checksum( newid())) % 100 
	from dbo.master

go 100 -- 100 executions of this batch

select *
from dbo.master m
inner join dbo.detail d
	on d.MasterID = m.MasterID
;

select
	MasterID, Name, StartDate, [Something], [Something else], [another thing]
from (
	select 
		m.MasterID
		, m.name
		, m.startdate
		, d.code
		, d.intvalue
	from dbo.master m
	inner join dbo.detail d
		on d.MasterID = m.MasterID
	) p
	pivot (
		avg( p.intvalue )
	for code in (
		[Something]
		, [Something else]
		, [another thing]
		)
	) as pvt
;

Open in new window

http://msdn.microsoft.com/en-us/library/ms177410(v=sql.105).aspx

HTH
  David
0
 
Kevin CrossChief Technology OfficerCommented:
An alternative would be to rank the code-value pairs based on the foreign key to the master table. Then derived tables, select codes and values independently with the rank as suffix. You can join the pivots to get code, value, code, value, ... this approach works with conditional aggregates versus pivots also if there is a fixed maximum number of codes/values per master record.
0

Featured Post

Get quick recovery of individual SharePoint items

Free tool – Veeam Explorer for Microsoft SharePoint, enables fast, easy restores of SharePoint sites, documents, libraries and lists — all with no agents to manage and no additional licenses to buy.

Tackle projects and never again get stuck behind a technical roadblock.
Join Now