intnan 0.1.4

Function collection for handling integers with NaNs

Integer data types lack special values for -inf, inf and NaN. Especially NaN as an indication for missing data would be useful in many scientific contexts.

Of course there is numpy.ma.MaskedArray around for the very same reason. Nevertheless, it might sometimes be annoying to carry a separate mask array around. And in those cases, using a set of numpy-compatible functions for the same job will do just fine.

This package provides such an implementation for several standard numpy functions, that treat integer arrays in such a way, that the lowest negative integer resembles NaN.