
Hi everyone,
I'm encountering a strange issue and could use some guidance on how to troubleshoot it.
I have a Tableau dashboard using two data sources:
- Tableau Prep Output – This connects to several Redshift tables and outputs key fields like vehicle number, batch number and several additional columns.
- Presto/Trino SQL Source – This contains billions of records and provides metrics per vehicle, including mileage, health, and warranty.
- I am not blending them for now.
Goal:
I want to create a dashboard where selecting a batch number filters the related
vehicle numbers , which then filters the Trino data source to show the corresponding vehicle-level metrics.
Problem:
The filtering doesn’t work as expected when using filters or dashboard actions—it shows incorrect values from the Trino source, and sometimes I can’t even tell what values are being pulled or why.
Observation:
It does work correctly when I pass the vehicle number as a parameter to the Trino source instead of using cross-data-source filtering. However, that approach isn't ideal for interactive filtering or multiple selections, because users have to manually view the vehicle number from a dropdown and then enter it into a vehicle search parameter, just like they do for the batch search.
Has anyone run into something similar? Any suggestions on what might be going wrong or best practices for handling this type of setup? I’ve attached an example workbook to illustrate the structure (though the data is limited).
Thanks in advance for any insights!

A workaround exists using the (so called) 'Header-as-Button' technique: