Closed mbostock closed 3 years ago
No problem in principle, but I believe this might break a lot of applications that rely (knowingly or not) on null being treated as 0. And I don't see a way to evaluate the risk (eg by searching for examples in blocks or observable).
I am confident we can proceed with this as a patch release. I think it is more likely that null data is inaccurately being misrepresented as zero than people are relying on null implicitly representing zero.
I’m surprised we didn’t find this sooner, but I think it’s appropriate for quantitative scales to treat null as undefined, rather than implicitly coercing null to zero. This makes d3-scale more consistent with d3-array.