Magic ETL - Allow group by without a grouping column

GrantSmith
GrantSmith Indiana πŸ₯·
in Ideas Exchange

There are times when I want to calculate aggregated metrics across the entire dataset. Currently this isn't possible without adding a constant value column to group on which I then have to remove later in my pipeline. It'd be much smoother if the Group By tile allowed us to group across the entire dataset without having to use a bogus constant column.



**Was this post helpful? Click Agree or Like below**

**Did this solve your problem? Accept it as a solution!**
Tagged:
9
Up
9 votes

Submitted Β· Last Updated

Comments

  • ChristianW
    ChristianW βšͺ️

    So what are you saying is that you just need a grand total? Maybe you can share an example of what you are trying to achieve

  • GrantSmith
    GrantSmith Indiana πŸ₯·

    @ChristianW - Yes exactly what I'm looking to do but you can't get it without adding a constant value to group on in the group by tile. SQL and Window Functions don't require a column to perform a grand total.



    **Was this post helpful? Click Agree or Like below**

    **Did this solve your problem? Accept it as a solution!**
  • ChristianW
    ChristianW βšͺ️

    Sorry, but I am not sure if you can just add a grand total within the ETL, probably that would be the best to do in a card? However, since you mentioned window functions you can do the same with the Rank and Window tile:

    Set the function as a Sum, select the column which contains your values and set it to unbounded to 0

    aa.PNG








    Then you can set the order and partitions if you wish

    bb.PNG


Categories