So Trifacta (CDP) can either append, replace or create new file each time the Job is run.
You're refering to an Upsert\Merge action.
So this can be done by 2 steps.
- for new rows, you can use the "append" run option
- for updating existing rows - 2 options.
Either -
A. do a Union transformation with the output file (already in bigquery), then deduplicate\aggregate and choose how you'd like to merge between the duped rows. this then can be done each time a new file is made\created\updated.
B. create a new file, then through parameterization import all files within the folder - unioning them and do the grouping mentioned in A.
Obviously - option A seems more simplistic.
Does that make sense?