Does starting a new prodigy session with same dataset name for a different input file deletes the old data which was present in the database. Or does it add new records to the old database.
Hi! Datasets in Prodigy are append-only and completely independent from the input data source. Whatever you annotate gets saved to the given dataset when you save your annotations.
So when you re-start the server with a different input file and the same dataset, new examples in the input file are shown to you for annotation, and then added to the existing dataset. Examples that are already in the dataset are skipped, so you're not asked the same question twice.