Incremental database updates will now by default auto-prune to remove previously-indexed files that no longer exist. This can be disabled by passing prune=False to the indexing functions.
Additionally, we use a broader and hopefully more robust search to figure out what the time dimension is within a file.