One way to solve this is to use Python. Not that I am a Python expert, but ChatGPT can help with that. Attached working sample.

import fmeobjects

class FeatureProcessor(object):
    def __init__(self):
        self.values = []

    def input(self, feature):
        # Get the attribute and period values
        value = feature.getAttribute('value')
        period = feature.getAttribute('period')
        
        # Convert value to integer if possible
        try:
            value = int(value)
            self.values.append(value)
            
            # If we have at least as much values as the period is, calculate the rolling average
            if len(self.values) >= period:
                rolling_avg = sum(self.values[-period:]) / period
                feature.setAttribute('rolling_average', rolling_avg)
        except (TypeError, ValueError):
            # Handle cases where conversion to integer fails
            feature.setAttribute('rolling_average', None)
        
        # Output the feature
        self.pyoutput(feature)

    def close(self):
        pass

Hi ​@bromp 

The only option shown here is to use a user parameter… but is it possible to assign an attribute to a user parameter that could vary with each run?

I don’t know if this is possible another way (like ​@nielsgerrits  suggests to use in python) but it seems to me like this would be a great suggestion in ideas: To be able to use an attribute for a variable value in the Adjacent Features function.

Thanks ​@nielsgerrits and ​@crutledge ,

I have implemented the python solution and that works nicely, I’ll also submit an idea to allow the attribute selection in the adjacent features section.