[netCDFJava #RXP-917865]: aggregation

> John-
> 
> GrADS has a notion of aggregation using templates of file names.  I'm
> wondering about the best way to handle multiple files.  Should I
> generate RandomAccessFiles for all of them right at the start and switch
> to the appropriate one when reading data, or open each one when I need
> it.  Is there overhead to opening them all up front?   The advantage I
> see there is being able to throw an IOException if one of the files
> doesn't exist.
> 
> What do you do in NcML aggregation?

we make sure the files exists, but dont open them. open and close as needed. 
that scales to arbitrary # of files, and avoids overhead, since you may not 
need them.



Ticket Details
===================
Ticket ID: RXP-917865
Department: Support netCDF Java
Priority: Normal
Status: Closed