That is a rather interesting feed design. It seems to be exactly the same situation as the question you linked - the short is being used to change the antenna's resonant frequency. This is very necessary for 870 MHz as it is very low for an antenna that small. The wavelength of 870 MHz is 34.5 cm (30 cm at 1 GHz / 0.87 GHz), making a standard quarter wave antenna 8.6 cm long. This antenna is 1cm long, necessitating some clever design. As for exactly why it does that; I am not an antennas expert so I probably can't give you a very good answer. Also, their datasheet does not shed much light on how the antenna is constructed. The picture seems to show a meandering trace on the side of the antenna, but without more detailed information on how the antenna is constructed and oriented it is a little difficult to speculate on exactly what is going on.
When it comes down to small antennas like this, you can't think in terms of currents and voltages, you have to think in terms of electromagnetic waves. An EM wave from the transmitter will be guided down the waveguide and coupled into the antenna, more or less ignoring the short. It does make a good deal of difference what direction the wave is coming from and exactly where it gets coupled into the antenna as this will affect both the mode excited as well as the impedance seen at the input. Well, the short will affect it but it won't stop it. It still gets coupled into the antenna. The shorting pin affects how the EM fields inside the antenna oscillate, helping the antenna achieve resonance at 870 MHz.
Speaking of impedance, I presume you're going to build a matching network to properly match the antenna's complex impedance (72-j23 to 86-j5 ohms, according to the datasheet) to your 50 ohm trace?