Glad to hear you're playing with the sample data! The (0.0,0.0,0.0) xyz values are all false negatives in the data set (theoretically there's always an object somewhere...). As you've noticed, currently, the sensor's minimum range is 0.8 m (soon to be 0.25 m), and the maximum range is ca. 105 m. Since the sensor has no visibility under 0.8 m or beyond the maximum range, the sensor cannot identify where the "true positive" lies.
So this is all to say that, unfortunately, when points in a single frame have a (0,0,0) value, there is no way to know if the true return is below the minimum range, or beyond the maximum range.