rebuild_history() Efficiency

[Christian Sibo]

Currently, the rebuild_history function runs back to front (that is, oldest to newest data), replacing data as new items are encountered.

It would be more efficient, especially as the database grows over time, for the function to run front to back. That way, it would run until it either filled every column (being the sum of latest data) or ran out of rows.

This would prevent situations where the first n rows might be completely overwritten with newer data by the time the function ends, which resulted in wasted work.