How to write a only integers numpy 2D array on a txt file

Question

I know how to write the txt file using numpy.savetxt() but I can't get it to write the file using just integers. I have the following code:

new_picks = new_picks.astype(int)
np.savetxt(fname='newPicksData.txt', X=new_picks.astype(int))

This is how the matrix that I'm getting looks like:

2.900000000000000000e+01 3.290000000000000000e+02 1.000000000000000000e+00
4.300000000000000000e+01 1.080000000000000000e+02 1.000000000000000000e+00
4.300000000000000000e+01 1.950000000000000000e+02 1.000000000000000000e+00
5.600000000000000000e+01 1.510000000000000000e+02 1.000000000000000000e+00
5.600000000000000000e+01 9.700000000000000000e+01 1.000000000000000000e+00
7.000000000000000000e+01 2.840000000000000000e+02 1.000000000000000000e+00
3.500000000000000000e+01 3.170000000000000000e+02 1.000000000000000000e+00
5.400000000000000000e+01 2.110000000000000000e+02 1.000000000000000000e+00
6.400000000000000000e+01 1.180000000000000000e+02 1.000000000000000000e+00
5.400000000000000000e+01 3.700000000000000000e+01 1.000000000000000000e+00
1.300000000000000000e+01 1.950000000000000000e+02 1.000000000000000000e+00
1.300000000000000000e+01 1.680000000000000000e+02 1.000000000000000000e+00
1.300000000000000000e+01 2.780000000000000000e+02 1.000000000000000000e+00
4.900000000000000000e+01 2.200000000000000000e+01 1.000000000000000000e+00
4.900000000000000000e+01 1.040000000000000000e+02 1.000000000000000000e+00
4.900000000000000000e+01 7.500000000000000000e+01 1.000000000000000000e+00
5.400000000000000000e+01 2.610000000000000000e+02 1.000000000000000000e+00
5.400000000000000000e+01 2.600000000000000000e+02 1.000000000000000000e+00
5.400000000000000000e+01 1.150000000000000000e+02 1.000000000000000000e+00
5.400000000000000000e+01 5.400000000000000000e+01 1.000000000000000000e+00
1.300000000000000000e+01 5.400000000000000000e+01 1.000000000000000000e+00
4.900000000000000000e+01 5.400000000000000000e+01 1.000000000000000000e+00

What I'm looking for is something like this:

29 329 1
43 108 1
43 195 1
56 151 1
56 97 1

Answer 1

You have to add an extra parameter in

savetxt(fname='newPicksData.txt', X=new_picks.astype(int), fmt ='%.0f\n')

that is just the formating of the actual number.

Answer 2

You have to specify the format separate using the fmt argument, see the documentation here http://docs.scipy.org/doc/numpy/reference/generated/numpy.savetxt.html

You have to use the following syntax to get what you want:

np.savetxt(fname='newPicksData.txt', X=new_picks.astype(int),fmt="%i")

Omitting that argument, defaults it to fmt='%.18e', which is exactly what you see in the output that you posted in the question.

---

Furthermore, do you really need the astype(int) part in your code? The np.savetext command can format strings/doubles perfectly fine into int without changing the data yourself. This can be demonstrated with the below code chunk:

import numpy
x = numpy.array([1.3,2.1,3.9,4.2,5.5,6.1])
numpy.savetxt(fname='newPicksData.txt', X=x, fmt="%i")

The listed code chunk then yields the following text file:

1
2
3
4
5
6