Just upgraded numpy to the latest intel conda’s, I am at pandas 2.2.1, numpy 1.24.3 and numexpr 2.10.0
If I try something simple like
df=pd.DataFrame(index=[0,1,2,3])
df["A"] = [1,2,np.NaN, 3]
print(df)
then I get six run time warnings.
RuntimeWarning: invalid value encountered in less
has_small_values = ((abs_vals < 10 ** (-self.digits)) & (abs_vals > 0)).any()
RuntimeWarning: invalid value encountered in greater
has_small_values = ((abs_vals < 10 ** (-self.digits)) & (abs_vals > 0)).any()
RuntimeWarning: invalid value encountered in greater
has_large_values = (abs_vals > 1e6).any()
RuntimeWarning: invalid value encountered in less
has_small_values = ((abs_vals < 10 ** (-self.digits)) & (abs_vals > 0)).any()
RuntimeWarning: invalid value encountered in greater
has_small_values = ((abs_vals < 10 ** (-self.digits)) & (abs_vals > 0)).any()
For a more complicated code the result is just overwhelming, hundreds and hundreds of such messages.
Did numpy drop support for NaN ??
3