numPy gives nan while reading a negative number from a file

问题

I tried to read the contents of a 3x3 matrix stored in a text file having the following values.

−30  10  20  
 10  40 −50  
 20 −50 −10

I used numpy.genfromtxt as follows but when it gave nan in place of negative data values.

data = np.genfromtxt('file.txt',delimiter=' ')
print(data)
print(type(data[0][0]))

But the datatype of the negative value still shows float64:

I also tried reading the values as a list and then converting to numpy array but that also didn't work.

Is there any other way I can read negative values as a numpy array from a file input?

回答1:

You've got U+2212 MINUS SIGN characters in your file, not the U+002D HYPHEN-MINUS people usually think of as the minus sign character. genfromtxt doesn't handle that character. Replace the minus sign characters with hyphen-minuses before trying to parse the data.

回答2:

Try this code instead:

data=pd.read_csv("data.txt",delimiter=",")
data = DataFrame(data)

来源：https://stackoverflow.com/questions/52241853/numpy-gives-nan-while-reading-a-negative-number-from-a-file