The common case is that you wouldn't combine them. In this case since your ETL dataset contains all of the same information + two additional columns you would just use this dataset for all of the cards, the other dataset would exist but only for the purpose of getting the untouched data into Domo, then the ETL would automatically run.
I can see how if you have a data limit this may be of concern to you, maybe someone else or your account manager has a recommendation?
Thanks for the clarification. Few more questions.
What do you mean by 'ETL dataset'? Is this the new dataset that I have created? If yes, then I presume based on your answer that this so called 'ETL dataset' will hold all columns from my previous Dataset + 2 additional columns after the split. Please correct me if I am wrong.
How this new dataset will get updated? Let say I add few more raws in my old dataset, then this new dataset will automatically be updated?
Yes sorry by ETL dataset I mean the new dataset that you chose to create at the end of the ETL process (Input block-->ETL transforms-->Output block) where the output block is where you created a new dataset. This is what I'm referring to as the ETL dataset.
In the ETL settings just choose the option which will automatically run the ETL everytime your original (input) dataset gets updated. That way you don't have to worry about it. Alternatively you can set it to run on a schedule but having it run whenever the input dataset gets updated is my go to option to make sure things are in sync.