문제
Here's a MWE of my code:
https://stackoverflow.com/questions/18498560
italiano
english
français
española
中国
日本の
العربية
Deutsch
한국어
Português
Russianimport numpy as np
# Load data from file.
data = np.genfromtxt('data_input', dtype=None, unpack=True)
print data
Here's a sample of the data_input file:
01_500_aa_1000 990.0 990.0 112.5 0.2 72 0 0 1 0 0 0 0 0 0 0 0 0 1
02_500_aa_0950 990.0 990.0 112.5 0.2 77 0 0 1 0 0 0 0 0 0 0 0 0 1
03_500_aa_0600 990.0 990.0 112.5 0.18 84 0 0 1 0 0 0 0 0 0 0 0 0 1
04_500_aa_0700 990.0 990.0 112.5 0.18 84 0 0 1 0 0 0 0 0 0 0 0 0 1
The unpack argument does not appear to work since it always prints:
[ ('01_500_aa_1000', 990.0, 990.0, 112.5, 0.2, 72, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1)
('02_500_aa_0950', 990.0, 990.0, 112.5, 0.2, 77, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1)
('03_500_aa_0600', 990.0, 990.0, 112.5, 0.18, 84, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1)
('04_500_aa_0700', 990.0, 990.0, 112.5, 0.18, 84, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1)]
Can anybody reproduce this? What am I doing wrong?
해결책 4
I'm posting my own answer since this is what I ended up using.
import numpy as np
# Load data from file.
data = np.genfromtxt('data_input', dtype=None)
# Force transpose list.
data = zip(*data)
This actually works and it's pretty easy to understand and use.
다른 팁
You're getting this because genfromtxt is returning a numpy record array, not a list. It's just that when you print() it to the console it looks like a list.
from cStringIO import StringIO
raw = """01_500_aa_1000 990.0 990.0 112.5 0.2 72 0 0 1 0 0 0 0 0 0 0 0 0 1
02_500_aa_0950 990.0 990.0 112.5 0.2 77 0 0 1 0 0 0 0 0 0 0 0 0 1
03_500_aa_0600 990.0 990.0 112.5 0.18 84 0 0 1 0 0 0 0 0 0 0 0 0 1
04_500_aa_0700 990.0 990.0 112.5 0.18 84 0 0 1 0 0 0 0 0 0 0 0 0 1"""
sio = StringIO(raw)
data = genfromtxt(sio, dtype=None, unpack=False)
print data
print
print data.dtype
gives:
[ ('01_500_aa_1000', 990.0, 990.0, 112.5, 0.2, 72, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1)
('02_500_aa_0950', 990.0, 990.0, 112.5, 0.2, 77, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1)
('03_500_aa_0600', 990.0, 990.0, 112.5, 0.18, 84, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1)
('04_500_aa_0700', 990.0, 990.0, 112.5, 0.18, 84, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1)]
[('f0', 'S14'), ('f1', '<f8'), ('f2', '<f8'), ('f3', '<f8'), ('f4', '<f8'), ('f5', '<i8'), ('f6', '<i8'), ('f7', '<i8'), ('f8', '<i8'), ('f9', '<i8'), ('f10', '<i8'), ('f11', '<i8'), ('f12', '<i8'), ('f13', '<i8'), ('f14', '<i8'), ('f15', '<i8'), ('f16', '<i8'), ('f17', '<i8'), ('f18', '<i8')]
unpack=True and unpack=False appear to return the same thing because you need a recarray. I would suggest you try pandas and forget about recarrays altogether. You can pass a recarray to the pandas.DataFrame and actually get s*** done! For example,
df = DataFrame(data)
print df
print
print df.f0
yields:
f0 f1 f2 f3 f4 f5 f6 f7 f8 \
0 01_500_aa_1000 990.00 990.00 112.50 0.20 72 0 0 1
1 02_500_aa_0950 990.00 990.00 112.50 0.20 77 0 0 1
2 03_500_aa_0600 990.00 990.00 112.50 0.18 84 0 0 1
3 04_500_aa_0700 990.00 990.00 112.50 0.18 84 0 0 1
f9 f10 f11 f12 f13 f14 f15 f16 f17 f18
0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 1
2 0 0 0 0 0 0 0 0 0 1
3 0 0 0 0 0 0 0 0 0 1
0 01_500_aa_1000
1 02_500_aa_0950
2 03_500_aa_0600
3 04_500_aa_0700
Name: f0, dtype: object
As mentioned by @Phillip Cloud, you are getting a recarray as you have a mix of data types (strings and numbers) - the strings in column 0 are causing this.
You could get around this by importing column 0 separately:
>>> np.genfromtxt('data_input', usecols=range(1,18))
array([[ 9.90000000e+02, 9.90000000e+02, 1.12500000e+02,
2.00000000e-01, 7.20000000e+01, 0.00000000e+00,
0.00000000e+00, 1.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00],
[ 9.90000000e+02, 9.90000000e+02, 1.12500000e+02,
2.00000000e-01, 7.70000000e+01, 0.00000000e+00,
0.00000000e+00, 1.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00],
[ 9.90000000e+02, 9.90000000e+02, 1.12500000e+02,
1.80000000e-01, 8.40000000e+01, 0.00000000e+00,
0.00000000e+00, 1.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00],
[ 9.90000000e+02, 9.90000000e+02, 1.12500000e+02,
1.80000000e-01, 8.40000000e+01, 0.00000000e+00,
0.00000000e+00, 1.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00]])
>>> np.genfromtxt('data_input', usecols=0,dtype=None)
array(['01_500_aa_1000', '02_500_aa_0950', '03_500_aa_0600',
'04_500_aa_0700'],
dtype='|S14')
Or, you could reference the columns in the recarray like this:
>>> data['f0']
array(['01_500_aa_1000', '02_500_aa_0950', '03_500_aa_0600',
'04_500_aa_0700'],
dtype='|S14')
>>> data['f5']
array([72, 77, 84, 84])
I can reproduce this. However, if your change dtype to float I get
[[ nan nan nan nan]
[ 9.90000000e+02 9.90000000e+02 9.90000000e+02 9.90000000e+02]
[ 9.90000000e+02 9.90000000e+02 9.90000000e+02 9.90000000e+02]
[ 1.12500000e+02 1.12500000e+02 1.12500000e+02 1.12500000e+02]
[ 2.00000000e-01 2.00000000e-01 1.80000000e-01 1.80000000e-01]
[ 7.20000000e+01 7.70000000e+01 8.40000000e+01 8.40000000e+01]
[ 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00]
...
I got the idea from this mailing list question.
Look at an answer given here. np.genfromtxt() returns data of the type ndarray. This cannot be heterogeneous.
