Efficiency of Microvertex code(MuKpi)- Using Pure D0 sample

Is the code actually reconstructing all D0s?/How much is lost?