Skip to main content
Announcements
Accelerate Your Success: Fuel your data and AI journey with the right services, delivered by our experts. Learn More
cancel
Showing results for 
Search instead for 
Did you mean: 
richard24best
Creator II
Creator II

How to avoid Synthetic keys and Create mapping between 2 different tables

Hi Team,

I have 2 tables - 

1 )GFXSettAnalysis (Summary table)

2) GFX Settled (Detailed table)

Requirement is that i want to create filters using the fields shown in Synthetic table so that users when filter they can see Summary view and in Detailed view. Is this possible ?

Not sure how do i map these fields, please assist.

Regards,

Richard

Labels (1)
1 Solution

Accepted Solutions
marcus_sommer

In many cases it's not expedient to keep multiple fact-tables within the data-model because it could cause such issues with synthetic keys and/or circular references (by further connecting them to dimension-tables). Further the relationship between the tables and/or missing key-values on one or even both sides could make serious problems. More disadvantages occur with multiple different fields for the same information which leads to a lot of efforts to address them properly.

Therefore I suggest you follows the official recommendations to create a star-scheme data-model. This means merging all facts into a single table and in your case just concatenating both facts together. By adding an extra source-field you will be further able to differentiate between the different tables.

- Marcus

View solution in original post

3 Replies
Mark_Little
Luminary
Luminary

HI @richard24best 

Would help if you could share you fields names, Getting a synthetic key, means you have fields in both tables with the same name, if the contain the same data then you can remove them from one table or if they have the same name but different data then you want to alias in one table. i.e.

Field AS GFXSettled_Field,

richard24best
Creator II
Creator II
Author

Thanks for your response Mark.

yes the field names are same, should i comment in one of the table to avoid Synthetic keys?

Regards,

Richard

marcus_sommer

In many cases it's not expedient to keep multiple fact-tables within the data-model because it could cause such issues with synthetic keys and/or circular references (by further connecting them to dimension-tables). Further the relationship between the tables and/or missing key-values on one or even both sides could make serious problems. More disadvantages occur with multiple different fields for the same information which leads to a lot of efforts to address them properly.

Therefore I suggest you follows the official recommendations to create a star-scheme data-model. This means merging all facts into a single table and in your case just concatenating both facts together. By adding an extra source-field you will be further able to differentiate between the different tables.

- Marcus