This is a request that would probably require a re-architecting of the entire database schema, but I'll give it a shot anyway. The screen capture below shows a bad reading that I want to override in the database because it will screw up my time series statistics, chart scales, etc. However, if I override it by uploading a CSV with the "patch" value (including a security code, if I'm smart about it), I would also wipe out the source data in the database. I always have the fallback of going to the backup CSV that LoggerNet is storing, but that decouples eagle.io from the source data forever.
Of course, the example screen capture shows only one of many possible scenarios in which data may have to be overridden. It would be great if a data versioning feature could be introduced in eagle.io so that the entire history of any data point could be audited (with a code field, a comments field for explaining the reason for the change, and an optional hyperlink field to a troubleshooting or calibration report, etc.). Otherwise, the CSV (or whatever backup system one chooses uses), not eagle.io, is the true historical data repository. Perhaps eagle.io is not intended to be a complete data respository, but only a data store for "cleansed" data(?)
I suspect this would be a lot of work to implement, but thanks for hearing me out!