From a physics standpoint, you need to consider your full cycle.
If you increase pressure by pumping in more molecules to a given volume or take a given set of molecules and reduce the volume makes a difference. (Usually high pressure applications like this will end up on the given volume route due to space constraints.) Pressure losses over check valves can have more than an insignificant bearing on easily compressible fluids such as gasses. Compressing very slowly can be efficient but your industrial process will likely need to sacrifice a bit of efficiency to get a desired throughput.
Next up the actual removal of heat. The compression causes the oxygen to increase in temperature. Increased temperature means heat will naturally flow out into a reservoir of the pre-compression temperature, but this is usually too slow for industrial processes. A heat pump of some sort would incur the remaining losses in bringing the compressed oxygen down to condensation temp for your pressure and removing energy at the rate desired for your process (this is where that kWh calculation comes in, but with the energy you put in for compression also included!).
Put both processes together and optimize to find an efficient target pressure and temperature. (One determines the other from Oxygen's phase diagram).