can't write an edf with a signal including np.NaN values

[CyrilLeMat]

Hi,

I may misunderstand how to use the pyedflib library, but it seems impossible to write a edf file with missing values (which would reflect in np.NaN in a numpy array)

example:

from pyedflib import highlevel
signals = [np.array([1, 2, np.NaN, 3])]
highlevel.write_edf('test.edf', signals, highlevel.make_signal_headers(['ch1']), highlevel.make_header())

[skjerns]

Indeed, the edf standard does not allow missing values. You will need to replace them.

In the background, edf saves data as integers (digital amplifier values). Ints can't represent NaN.