I see. So this is a homework problem?
This is like asking for a general-purpose glue that will stick to anything, or a universal linkage that will work as any kind of lever. A real temperature sensor (even one with a very small thermal mass) will have a very s-l-o-w response time in electronic terms, and even a simple 1st-order filter would be enough to get rid of any real-world high frequency noise I can think of, and still retain the details of the signal. A current sensor is very different, the signal you are filtering out may be exactly what you need to be reading. You need to look at it. There is no general case.
For a low-pass filter like your example, there isn't much advantage to one over the other. The active filter will have a somewhat sharper cutoff (owing to a higher Q), and exhibits slightly less noise from the amp (owing to a higher signal at the input). The real advantages are in things like bandpass filters, where you can jack the Q up until it rings (if you want, I made a cool electronic drum set this way), give it a much higher input impedance, and you don't need inductors - which can easily cost more than the rest of the circuit combined.