I was wondering what the reason it is that when designing Type II storage areas that we have to group the tables with similar RPB into the same storage area and not consider their relationships instead? For example if I have invoice headers and invoice charges which will most likely always be read together, should I not put them in the same storage area despite them being quite different in record size?
Another question I have is with a particular table that has its mean record size inflated by some of the records having very large 'notes' field (basically at the 32KB limit). If I don't count those records then the mean record size is around 300-400 bytes but with the notes inflation, it pushes the mean record size to 1100+. Those are not anomalies and will happen every now and then. Should I size the RPB based on the mean size with the inflated record size or should I ignore those and size based on the average record size of the majority of the records since there will be fragmentation anyway?
Thanks in advance!
Another question I have is with a particular table that has its mean record size inflated by some of the records having very large 'notes' field (basically at the 32KB limit). If I don't count those records then the mean record size is around 300-400 bytes but with the notes inflation, it pushes the mean record size to 1100+. Those are not anomalies and will happen every now and then. Should I size the RPB based on the mean size with the inflated record size or should I ignore those and size based on the average record size of the majority of the records since there will be fragmentation anyway?
Thanks in advance!