We have a financial application that writes a large volume of data, of the order 100GB per day. We tag the data according to the symbol, and whereas data for a given symbol is always sorted in timestamp order, different symbols may not be. For instance, we might want to load 1 day of data for symbol A, followed by 1 day of data for symbol B.
Is there a performance penalty to loading such data out-of-order, given that it’s under separate tags?
