Real World SQL

Using Row Count data flow transformations

Jeff Rix — Thu, 19 Jan 2012 15:13:38 +0000

Create an Int32 user variable with a package scope. Configure the Row Count property Custom Properties>Variable Name to the user variable.

You cannot use the Row Count variable in the same dataflow it is set because it is processing the records and sending them to the destination at the same time.

to use in the variable in an Execute SQL Task use an ADO.Net Connection Type and map the variable to an input parameter with the same data type using a parameter name that starts with the @ character.

In your query reference the variable using the parameter name and you will be all set!

Duplicate Keys in a dimensional hierarchy

Jeff Rix — Fri, 13 Jan 2012 21:41:40 +0000

This date hierarchy has year at the highest level and date at the lowest level. Months and quarters are between the highest and lowest levels in this data hierarchy.

Each year has a fourth quarter and each Quarter has the month of December.

The default property for dimensions in SSAS are set up to treat this as an error.

For most dimensions this would align with the business rules. If you had a football

league roster you would want the dimension to throw an error if it had a player

playing on more than 1 team.

Change the Key Duplicate Error Configuration to Ignore Error.

MDX Where clause

Jeff Rix — Fri, 13 Jan 2012 21:21:53 +0000

This MDX query shows delinquent balance and original principal balance across the servicing plan members with a rate type id equal to 1.

SELECT {[measures].[Delinquent Balance], [measures].[Original Principal Balance]} ON COLUMNS ,NON EMPTY [Servicing Plan WID].[Servicing Plan ID].MEMBERS ON ROWS FROM [Cube T]
WHERE {([Dim Rate Type].[Rate Type ID].&[1])};

Update using joins- Query returning multiple values

Jeff Rix — Fri, 13 Jan 2012 15:33:23 +0000

If you are updating a tables column by selecting a column from another table there are different syntax you may use. If the select statement’s column contains more than 1 value for the specified join condition some syntax’s will update the value multiple times while another will fail the update statement.

In this example Dennis Rodman has multiple jersey player numbers in the UpdatePlayerData table.

create table PlayerData

Player_ID int identity(1,1) primary key,

Player_Name varchar(50) not null,

Player_Number int not null

insert into PlayerData values

('Michael Jordan', 23),('Scottie Pippen', 33),('Dennis Rodman', 91)

create table UpdatePlayerData

Player_ID int not null,

Player_Number int not null

insert into UpdatePlayerData values

(3, 10),(1, 45),(3,73)

The first two statements execute successfully and use the same execution plan. They update Dennis Rodman’s jersey to 10 and then to 73.

--Update 1

update PlayerData

set Player_Number = x.Player_Number

from

select Player_ID, Player_Number from UpdatePlayerData

) as x

where x.Player_ID = PlayerData.Player_ID

--Update 2

update a

set a.Player_Number = b.Player_Number

from PlayerData as a join

UpdatePlayerData as b on a.Player_ID = b.Player_ID

The last value for his jersey is what is now displayed in the table.

This update syntax will fail the query stating that the sub-query returned more than 1 value.

--Update 3 that will fail

update PlayerData

set Player_Number = (select upd.Player_Number from UpdatePlayerData as upd

					where upd.Player_ID = PlayerData.Player_ID)

where Player_ID in (select distinct Player_ID from UpdatePlayerData)

Even though it fails in this example I would generally advice using this syntax so you can control the logic that decides which value it will set the column to as opposed to just setting it to the last value it processes.