I'm looking through my local components distributor's online catalogue, and bridge rectifiers are significantly cheaper for a given current rating than individual diodes. I'm likely to need two or more diodes anyway, and it's okay for them to be connected at one side or the other, so I'd only be "wasting" half the bridge and not three quarters of it.
But: is it safe to "rectify" a DC input? My intent is to protect string(s) of solar panels from reverse power, so the individual diodes in the bridge would be forward-biased for hours at a time. Do hot-spots form in the junctions if the input frequency is too low? Presumably a 35A bridge has big junctions; if it's safe to pass those 35A through that big junction for 10ms at a time, every 20ms (50Hz AC input), to me it doesn't necessarily follow that the junction can safely handle 17.5A (or anything roughly half of 35A) continuously, for similar sorts of reasons that you can't just parallel BJTs without adding emitter resistors. I've just had a look at two random datasheets and there's no hint of any time dependence in the maximum average forward current. One example: the KBPC35xx series
Actually this question applies equally well to individual rectifier diodes, it's just that bridges have so much higher current ratings relative to the package size that I'm more suspicious in that context.