How to use python numpy.savetxt to write strings and float number to an ASCII file?

Question

I have a set of lists that contain both strings and float numbers, such as:

import numpy as num

NAMES  = num.array(['NAME_1', 'NAME_2', 'NAME_3'])
FLOATS = num.array([ 0.5    , 0.2     , 0.3     ])

DAT =  num.column_stack((NAMES, FLOATS))

I want to stack these two lists together and write them to a text file in the form of columns; therefore, I want to use numpy.savetxt (if possible) to do this.

num.savetxt('test.txt', DAT, delimiter=" ") 

When I do this, I get the following error:

>>> num.savetxt('test.txt', DAT, delimiter=" ") 
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Library/Python/2.7/site-packages/numpy-1.8.0.dev_9597b1f_20120920-py2.7-macosx-10.8-x86_64.egg/numpy/lib/npyio.py", line 1047, in savetxt
    fh.write(asbytes(format % tuple(row) + newline))
TypeError: float argument required, not numpy.string_

The ideal output file would look like:

NAME_1    0.5
NAME_2    0.2
NAME_3    0.3

How can I write both strings and float numbers to a text file, possibly avoiding using csv ( I want to make if readable for other people )? Is there another way of doing this instead of using numpy.savetxt?

Answer 1

You have to specify the format (fmt) of you data in savetxt, in this case as a string (%s):

num.savetxt('test.txt', DAT, delimiter=" ", fmt="%s") 

The default format is a float, that is the reason it was expecting a float instead of a string and explains the error message.

Answer 2

The currently accepted answer does not actually address the question, which asks how to save lists that contain both strings and float numbers. For completeness I provide a fully working example, which is based, with some modifications, on the link given in @joris comment.

import numpy as np

names  = np.array(['NAME_1', 'NAME_2', 'NAME_3'])
floats = np.array([ 0.1234 ,  0.5678 ,  0.9123 ])

ab = np.zeros(names.size, dtype=[('var1', 'U6'), ('var2', float)])
ab['var1'] = names
ab['var2'] = floats

np.savetxt('test.txt', ab, fmt="%10s %10.3f")

Update: This example also works properly in Python 3 by using the 'U6' Unicode string dtype, when creating the ab structured array, instead of the 'S6' byte string. The latter dtype would work in Python 2.7, but would write strings like b'NAME_1' in Python 3.