SSAS：如何设计我的多维数据集

Question

I'm quite new to SSAS and have problem designing a cube.

• table 1: closed tickets with: ticket_id, user_id, date, processing_time
• table 2: user data: user_id, login, first_name, last_name
• table 3: user groups: group_id, group_name
• table 4: period of user belong to a group: user_id, group_id, assigned_date, left_date
  I also added a time table

How do I design the cube to get the amount of tickets a user has closed while he was a member of a specified group? I have no clue how to tell SSAS to select the group when the selected date is between assigned_date and left_date.

Answer 1

You need to denormalize the data to allow this to be handled intuitively.

I would handle this in a transformation when populating your TicketsMEasure in either your datawarehouse or SSAS cube. Joining tickets with users with group members ship, using ticket.closeddatetime between group start/end as criteria. This will denormalize your tickets such that you can now have the group ID that the user was a member of when the ticket was closed

ClosedTicketsMeasure
TicketId, UserId, GroupId, DateKey

Now you can easily slice/dice on Groups/Users/Months/Weeks to get counts of the number of tickets closed in those combinations.

Answer 2

Using MSFT's SSAS with SSDT, you must follow the steps below:

• Create a connection to the database, you can do it by using the intelligent wizard when pressing right click at fist folder on solution explorer of your project.
• Create a database model that connects your data, you can do it by using the intelligent wizard when pressing right click at second folder on solution explorer of your project, here select your connection manager that you built in previous step and the tables that you will use.
• Create a cube, you can do it by using the intelligent wizard when pressing right click at third folder on solution explorer of your project, here select that model you built in previous step and the measures you want to use.

After you do there's steps and make the necessary hierarchies and partitioning for efficiency improvement then you can deploy your cube on your server by pressing the start button and use the browser of SSDT using drag and drop objects like dimensions and measures to built your report.