High current relays (or, more properly, contactors) typically have two stationary contacts bridged by a moving shorting bar with two contacts. That's because making a flexure or using a flexible wire capable of high current is fraught with difficulties.
High voltage DC contactors have an additional difficulty- the contacts tend to arc, especially if there is a bit of stray inductance in the circuit, and the arc can be sustained for some time and erode the contacts in a very few operations unless the contactor is made very large, and even then the arcing will continue until the contacts are far apart. (AC arcs tend to extinguish near the zero crossings, depending on how much stuff has been ionized and flung into the air, but in any case, easier than DC arcs).
A few ways have been used to mitigate this, such as pressurizing the switching mechanism with \$\text {SF}_6\$ (sulfur hexafluoride). As well as being troublesome and expensive, that gas is a powerful greenhouse gas (tens of thousands of times more than CO2) so it's only used where it has to be. Another way is to use permanent or electromagnets as "blow out magnets" to put a physical force on the ionized arc stream, extending the length and extinguishing it faster.
A good description can be found on the Durakool (of mercury contactor fame) website:

The purple cylinders represent the contacts. A DC current is flowing down on the left and up on the right. Two magnets are arranged to blow the arcs outward, away from the shorting bar with force F. Note that the 'hand' illustrated is the left-hand rule.

Using the magnets allows high-voltage DC contactors and relays to be manufactured that are small and relatively inexpensive.
As the datasheet indicates, reversing the polarity can be worse than no magnets at all, and it's only rated for a a fraction of the current and much fewer operations (as low as 1 operation life for fault current).