Junction numbering for EPANET

Can someone please clarify the following for me on the input file for EPANET?

Do junction (or node) numbers have to begin with the number "1" and increment sequentially, or can any arbitrary number be assigned to a junction as long as it is an integer, not duplicative of another junction number and between 1 and 2147483647?

Thanks.
Morris

You are technically correct. Node numbers are of the type "typedef long INT4;" I found that, for practical applications, 5 digit node (and other ID) numbers were sufficient, therefore wrote EPADATA with 5 digit ID numbers.

Istvan

Morris,

Junction numbers (or any id number) need not be sequential in epanet. We use ranges of numbers to identify different types of features. e.g. 1000's are junction nodes, 2,000's are wells, 3000's are water treatment plants and so on. The same approach can be used with pipes, pumps and valves.

I don't know if there are any memory or performance impacts of non-sequential (sparse) numbering schemes. If there are they haven't had any practical impact on us.

I understand that when Lew Rossman releases the next version you will be able to use alphanumeric symbols to represent features. That will make models a lot more "readable".

Regards,
Ben Harding
blh@hydrosphere.com