I'm running into a consistent problem when trying to create table objects that I think is caused by a misunderstanding on my part about how tables use dimensions and measures.
If I have a table that should display all records from the data model respective to the user's filters, my understanding of the best case scenario is to have a PK as a dimension and place the rest of the columns in as ="" measures. However, my current data set does not have a PK, it is just a collection of rows and columns. My first thought was to simply add every field as a dimension, but this resulted in the columns displaying mostly blank records except for the fields that have unused data points. This throws my count expressions off.
My second thought was to try to parcel out what I could use to group data, but I am consistently coming up with issues where data is appearing as null where it shouldn't.
Is there documentation on this, or could someone help me with an explanation? This seems like it should be simple, but I'm spending more time trying to deconflict tables like this than I should be.