If your talking about someone clapping with the beat of the music, I'd use spectral repair. Within that module, it really depends on how big (ie. how loud and how audible the overtones are) on which setting I'd use. The small stuff (like removing a close conversation) is usually best done by breaking to down to it's smallest components (syllables) and using the replace setting. This really can take some time, but if you try and remove too big a chunk, you'll get some of the music for sure. The next step up (which would include individual claps I think) would be to use the pattern setting. You have to be careful with the settings or you'll get echoing from the areas surrounding the bit you're trying to get rid of. Sometimes, that's impossible and I'll go back and use the attenuate setting on the echo to minimize it. Sometimes, if the frequency is pretty steady and not too wide, you can use replace and do it in several bits. But even with pattern, it's better to break it down into several bit (that's how I do close coughs) and then maybe go after the overtones in a couple of big chunks.
For the most part, just like panther65 said, it takes a lot of patience. I'm still at the trial and error point so don't take my word as fact. This is just what I've been doing so far. Anything you come up with would be very welcome.