Return a p-value of 0 for ER SKATExactBin_Work when we have no data to act on. Works around a crash in efficient resampling.
I don't know enough about the method to know if this is the right solution. Maybe we want to return NaN, maybe we want to return infinity, maybe we want to throw an exception and try a different approach entirely? We've got a dataset that fails here because p1 is empty and thus we cannot take the mean of it: https://github.com/saigegit/SAIGE/issues/88
@weizhouUMICH @saigegit
Return a p-value of 0 for ER SKATExactBin_Work when we have no data to act on. Works around a crash in efficient resampling.
I don't know enough about the method to know if this is the right solution. Maybe we want to return NaN, maybe we want to return infinity, maybe we want to throw an exception and try a different approach entirely? We've got a dataset that fails here because p1 is empty and thus we cannot take the mean of it: https://github.com/saigegit/SAIGE/issues/88