Under some conditions genfromtxt seems to mask not only missing data, but also data that happens to have the value as the filling_values (or its default if no filling_values is given).
While setting the filling_values to an "unlikely" value circumvents this i still find it troubling. Oddly, it only appears if I read more than one column.
In the attached example code the -1 (which is the default filling value for integers) is masked unnecessarily.
Original ticket http://projects.scipy.org/numpy/ticket/2158 on 2012-06-13 by trac user hoesterlin, assigned to unknown.
Under some conditions genfromtxt seems to mask not only missing data, but also data that happens to have the value as the filling_values (or its default if no filling_values is given). While setting the filling_values to an "unlikely" value circumvents this i still find it troubling. Oddly, it only appears if I read more than one column. In the attached example code the -1 (which is the default filling value for integers) is masked unnecessarily.