We discussed this some at Omniture Summit 2011 but wanted to submit this officially as an idea for discussion. When changing dates for datasets, that necessitates the dataset to be completely dropped and reprocessed. Most of our datasets are now a rolling 13 months of data. None of our datasets have an end date. When it's time to archive the oldest month, we need to change the start date of the dataset and it re-processes from the beginning. Is there a means where the start date could be modified without causing the entire dataset to reprocess? If we're just archiving the oldest month's vsl files, albeit visitor buckets are affected, but the other data is accurate and could be available for querying. Thoughts?
