Hi everyone,
We have a script that receives an input and writes a Tableau hyper file and consequently publishes in Tableau using our own Tool (developed in Python).
Everything works perfectly, except when the dataset is empty (which we still need it in Tableau for Tableau to function properly).
As you can see below we are using the Block Until Done so we wait for the .hyper file to be written on disk to then publish the file in Tableau.
Note that the below part under the container is just a script to get the file path and load it in our own tool (ignoring the input dataset which will be by then written in file).

The strange situation when the dataset is empty is that the Tableau Publish fails saying that the file doesn't exist (but it does exist). If we rerun the workflow with the exact same settings (and empty dataset) it will now publish to Tableau without issue, because the file already exists (from the previous run).
It seems like somehow, because the dataset is empty the block until done doesn't wait for the file to be written.
Anyone has some insights on similar situation?
Thanks,
Bruno