Network dataset - Costs: How to create a evaluator expression referencing a field of the feature class

[ConnieSu]

Hello! I have a network dataset and I want to create a evaluator expression in "Costs" referencing a field of the feature class. Can somebody help me, please?

These are my current costs of the ND.

For the edges feature class "PNL2017_Rodovias_Completo", I would like to create a expression like this:

if(TYPENO=="7001"):
    return 0.069001*[Shape]
else:
    return 0.039429*[Shape]

 And for the junctions feature class "PNL2017_Suporte_Pontos_Rodo", I would like something like this:

if(TYPENO=="7001"):
    return 17.1361
else:
    return 35.18

Could you please help me? 

Thanks and regards!

[MelindaMorang · Accepted solution]

You can do this with a field script evaluator.  Those evaluator dialogs are a little hard to use, though, so I can see why you need a hint!

First, for the evaluator you want to update, double click in the Value column until you see the box on the far right with the green X.

You'll get a pop-up box where you can create an expression for the field script evaluator.  You will want something like this:

Basically the Result box calls a custom-defined function called get_value() and passes the TYPENO and Shape field values to it.  The ! surrounding the field names is what's used for Python evaluators.  I'm using Python instead of VBScript because it's easier to code in (for me), and VBScript evaluators will be deprecated soon.

Here's the code for the codeblock:

def get_value(typeno, shape):
    if typeno == "7001":
        return 0.069001 * shape
    else:
        return 0.039429 * shape

Your expression for the junctions source will be very similar, but you won't need to pass the Shape field into the function.

[ConnieSu]

Thanks for your reply Melinda! It worked!

I've read that VBScript processes faster than Python. Will Python be faster when VBScript is deprecated?

Thanks and regards!

[MelindaMorang]

For field script evaluators like these, all the processing is done when you call Build Network.  The values are stored in the network dataset and are just looked up at solve time, so there shouldn't be any performance difference between Python and VBScript evaluators.  If there is, that performance difference would be incurred at build time, not at solve time.

However, I believe the performance difference is mostly historical at this point.  There certainly was some difference back in the ArcMap days, but I think that is either gone or significantly less now, although I don't have specific data to back that up.

Note that element script evaluators are different from field script evaluators.  Their code is run at solve time, so the solve may be slower, depending on the efficiency of the code itself.  This does not apply to your situation.