wrong reduced_chi_square

[sfzastro]

Hi,

I use 31 filters in my fitting, and not all objects have so many valid fluxes.
However, in the fitting results, the best.chi_square / best.reduced_chi_square is 30 for all my objects.

CIGALE version: 2022.1 - Python version: 3.8.12 - Platform: linux-x86_64

[Guang91]

Have you followed the figure for invalid fluxes/errors? (the figure is from the CIGALE manual)

[sfzastro]

I set the invalid fluxes to -99, which no longer works.
Thanks.

[mboquien]

The easiest way is to put flux and uncertainties to NaN. That way, there is no ambiguity.