SSIS 2008 : Need some advice where source is Type 1 SCD and stores as Type 2 SCD

Hi All

Points for any constructive advise on an SSIS project I have where the source is a Type 1 SCD, and desired target is SQL Server that stores the data as Type 2 SCD.

Source
Salesforce.com (SF) data, PK’s are columns id (int) and SystemModstamp (datetime) columns, and the data is stored as a Type 1 SCD, which means If a value changes in a row then it will have the same id, but different SystemModstamp, and whatever other changes.  No history is kept.

My Requirements
Client wants see…
•      (Requirement #1, aka R1) Only the current row (i.e. same as source)
•      (R2) All history for a given row, which means I have to build that up on my own based on changes over time.

My Table / Views Design
•      ‘Staging’ table as an initial landing for all source data.  No problems on this one.
•      A single table for each SF table, with all rows (to include changes as new rows),   Also includes a CurrentRow (bit) column with a value of 1 if current row and 0 if not current.
•      (R1) will be a view that is a SELECT * FROM above table
•      (R2) will be a view that is a SELECT * FROM above table WHERE current_row = 1
•      Indexed on id, SystemModstamp, and CurrentRow.

My SSIS Package Design
•      Truncate ‘staging’ tables, load with SF source data.
•      Then in a single data flow…
o      INSERT all rows (with current_rows = 1) without a match in id
o      INSERT all rows (with current rows = 0) with a match in id, and no match in SystemModstamp.
o      Run a SQL Stored Procedure to look at the rows with multiple SystemModstamp values for a given id, update the max to 1, all else 0.   This I can code, so no problem.

Again, points for any constructive feedback.

Thanks.
Jim
LVL 67
Jim HornMicrosoft SQL Server Developer, Architect, and AuthorAsked:
Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

Phillip BurtonDirector, Practice Manager and Computing ConsultantCommented:
I would have a bit more auditing.

Have Start Date and End Date as columns for your SCD 2, so you can trace where a row changes. I know you have the SystemModstamp, but it might make things easier - your call. Also, a column as to when a row was imported, so if your source data has an error, you can trace all rows entered then.

Do you need a new primary key for the SCD 2. And should it be non-clustered, so you can use your index as a clustered index?

What is the end result going to be used for? If mostly used for queries where CurrentRow = 1, then you might want the index to be on CurrentRow, then id etc.

Just some thoughts.
0

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
Jim HornMicrosoft SQL Server Developer, Architect, and AuthorAuthor Commented:
>Have Start Date and End Date as columns for your SCD 2, so you can trace where a row changes
Thanks.  Just dawned on me that instead of doing the WHERE CurrentRow = 1, I can always do WHERE SystemModstamp_End IS NULL and get the same result without all of the mechanics of populating a new column CurrentRow.

>Do you need a new primary key for the SCD 2.
id + Systemmodstamp will work fine as the PK.

>And should it be non-clustered, so you can use your index as a clustered index?
Probably could, as all SystemModstamp values will be data entry as of that day.  
Haven't thought about CLUSTERED vs. NON-CLUSTERED.

>What is the end result going to be used for?
All the current downstream reporting matches the source (my role is to add history to the current process), so yes an index on that value would be ideal.
0
Jim HornMicrosoft SQL Server Developer, Architect, and AuthorAuthor Commented:
Thanks.
0
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Microsoft SQL Server

From novice to tech pro — start learning today.