How's this for an idea?
1. Set up a primary data source that has all the dimensions.
2. Add both your data sources as secondary data sources using Tableau's data blending.
3. Add a quick filter that supports excludes on the primary data source.
I've set this up in the attached using the Superstore Sales data. Blending can introduce some complications in terms of aggregation and level of detail, and there are usually workarounds.
blend for filter.twbx.zip 1.1 MB
Yes, I thought about Blending... but it honestly scares me as I don't understand it well. What I always seem to run into is the issure around aggregations. But I think you may be on to something here. I will give it a try. You always seem to have some great ideas. Thanks so much!!!!
You're welcome! It's the only thing that I could come up with to meet all your requirements, besides doing work on the back-end to link your data sources.
I think I'm trying to do a similar thing, and I don't understand how this solution addresses the problem. If I have to create a primary source with all the data in it, why would I bother to then add them as secondary data sources also?
I have a primary data source (source 1) which gives me numerical and textual data for a number of items each of which has a unique index (primary key). I then have a secondary data source (source 2) which contains a list of items of interest - a subset of the master list in the primary source. But it is just the item primary keys. So I want to filter the details of items from the primary source to show only the items of interest listed in the secondary source.
When I try to do this using a filter, it has no effect because the dimension from source 2 has a different name from the dimension in source 1, even though I have told Tableau that they are linked by that dimension.
I tried creating a Boolean which tests the two dimensions for equality, so I can filter using that, but the calculated field insists I use an aggregation - which won't work (and keeps returning Null).