They could ALL be correct.
If you are measuring compressor motor starting current, initial starting current varies greatly depending on coolant temperature, pressure, position of pump, etc, at the time of measurement. Also, unless you are using calibrated meter, they are close enough, that I would simply read it as 20 Amp for 100 ms.
Keep in mind, in rush current is not a square wave. It's not that 20 amp is consumed for the whole 100 ms. So depending on how it is sensed and threshold is set, they are (again) close enough for practical use. If I were doing what you are doing, and it is critical, I would use current sensing probe along with an oscilloscope. Or, go further and use a shunt and measure the voltage, but that's going quite far, unless there is a very specific need.
For rest of your questions, the best thing you could do is to read off the manufacturer and part number off the motor and look for spec sheets. I'm sure they test that. If it's public information or not is a big guess.