|
In answer to your first query, you will probably find that you get
optimum performance by organising your documents into folders according
to what suits the customer rather than paying particular attention to
performance. This will apply to both documents in shared drawers as
well as private drawers.
Drawers are indexed in the DOCDB by folder name and document number.
This means that when a user indexes a drawer by folder only, then the
records corresponding to each document are got sequentially whereas if
he indexed by title documents in different folders, then he may get
records from different parts of the DOCDB leading to more I/Os. If he
indexed by author for example he may have to get every record in the
DOCDB as author is not a key.
Therefore by logically grouping documents into folders as the user
would logically group them they are more likely to be acccessed in the
most performant manner.
I can't answer your second question.
Mark
|