numpy.genfromtxt is not unpacking

问题

I'm having a strange issue with the package numpy.genfromtxt. I use it to read a data file with a number of columns (available here) but these are not being unpacked even when unpack is set to True.

Here's a MWE:

import numpy as np
f_data = np.genfromtxt('file.dat', dtype=None, unpack=True)

print f_data[3]
(237, 304.172, 2017.48, 15.982, 0.005, 0.889, 0.006, -2.567, 0.004, 1.205, 0.006)

(I use dtype=None because the file can have strings scattered around)

As you can see it returns a line instead of an unpacked column.

If I use np.loadtxt it works as expected:

f_data = np.loadtxt('file.dat', unpack=True)

print f_data[3]
[ 16.335  16.311  15.674  15.982  16.439  15.903  15.313  18.35   15.643  14.081  16.578  11.477]

What am I doing wrong here?

回答1:

Is this what you want?

In [448]: i=3
     ...: d=np.genfromtxt(fname, None) #d is a recorded array (or structured array)
     ...: d['f%d'%i] #Addressing Array Columns by Name
Out[448]: array([ 16.335,  16.311,  15.674,  15.982,  16.439,  15.903])

see:

http://wiki.scipy.org/Cookbook/Recarray

http://docs.scipy.org/doc/numpy/user/basics.rec.html#module-numpy.doc.structured_arrays

EDIT:

I tested d=np.genfromtxt('a.x', dtype=None, unpack=True) on the following data:

144     a578.06 873.72  16.335  0.003 
#-------^--------
180     593.41  665.748 16.311  0.003 
147     868.769 908.472 15.674  0.003
237     asdf.172 2017.48 15.982  0.005
#-------^--------

with dtype=None, unpack indeed fails:

In [538]: d=np.genfromtxt('a.x', dtype=None, unpack=True)
     ...: print d[3]
     ...: print d[1]
(237, 'asdf.172', 2017.48, 15.982, 0.005)
(180, '593.41', 665.748, 16.311, 0.003)

while with default dtype or dtype=str, unpack works:

In [539]: d=np.genfromtxt('a.x',  unpack=True)
     ...: print d[3]
     ...: print d[1]
[ 16.335  16.311  15.674  15.982  16.439  15.903]
[      nan   593.41    868.769       nan  1039.71    385.864]

In [540]: d=np.genfromtxt('a.x', dtype=str, unpack=True)
     ...: print d[3]
     ...: print d[1]
['16.335' '16.311' '15.674' '15.982' '16.439' '15.903']
['a578.06' '593.41' '868.769' 'asdf.172' '1039.71' '385.864']

回答2:

Change the

dtype=None

dtype=str

And remove unpacking, as this will transpose the data. And for good practice add delimiter :)

来源：https://stackoverflow.com/questions/21937979/numpy-genfromtxt-is-not-unpacking

标签

python

file-io

numpy