Negative IRR about pyxirr

[Anexen]

Hi @aluyca, it looks like the root finding algorithm found just one of several solutions to the equation.
An interesting point: LibreOffice shows the same result as pyxirr and mathematically it is correct because npv(pyxirr.irr(cf), cf) is close to zero and it's closer to zero than npv(npf.irr(cf), cf).
I'll see what I can do

[aluyca]

Thanks a lot!

[Anexen]

Hi @aluyca , I've tweaked IRR so that it first tries to find a rate greater than zero. I hope this helps. Please try version 0.7.2.

[aluyca]

Thanks!