unable to do np.savetxt on save newly appended string column to numpy array

Question

I have numpy array mfcc having mfcc values , and is of shape (5911,20).
I have one list a =[] which has 5911 labels like apple cow dog.
I want to append the these labels to the mfcc numpy array.

STEP1 I converted list with labels to an array :

at = np.array(a)
print (at)
print at.shape
print type(at)

['apple' 'apple' 'apple' ..., 'cow' 'cow' 'cow']
(5912,)
<type 'numpy.ndarray'>

STEP2 I made sure both at and mfcc were of same dimensions:

if len(at) > len(mfcc):
    at= at[ :-1]

STEP3 Then I stacked them together.

mfcc_with_labels=np.hstack((mfcc_with_labels,at[:,None]))
print mfcc_with_labels.shape 

(5911,21)

PROBLEM STEP Now I want to save this mfcc_with_labels to a file. So that I can feed it to a neural network later.

np.savetxt("mfcc_with_labels.txt", mfcc, newline= "\n", delimiter="/t")

and it throws a huge ERROR **

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-15-7709c644ca06> in <module>()
      1 print mfcc_features_with_times_and_labels.shape
      2 
----> 3  np.savetxt("mfcc_with_labels.txt", mfcc, newline= "\n", delimiter="/t")
/usr/local/lib/python2.7/dist-packages/numpy/lib/npyio.pyc in savetxt(fname, X, fmt, delimiter, newline, header, footer, comments)
   1256                     raise TypeError("Mismatch between array dtype ('%s') and "
   1257                                     "format specifier ('%s')"
-> 1258                                     % (str(X.dtype), format))
   1259         if len(footer) > 0:
   1260             footer = footer.replace('\n', '\n' + comments)
TypeError: Mismatch between array dtype ('|S32') and format specifier ('%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e/t%.18e')

**

I tried specifying 'fmt = %s' as an option but nothing happens.

I inspected and mfcc_with_labels[1] and the stacking/appending did work,

['-498.357912575' '-3.40930872496e-14' '1.55285010312e-14' '-5.31554105812e-14' '4.81736993039e-15' '-3.17281148841e-14' '5.24276966145e-15' '-3.58849635039e-14' '3.11248820963e-14' '-6.31521494552e-15' '1.96551267563e-14' '1.26848188878e-14' '6.53784651891e-14' '-3.15089835366e-14' '2.84134910594e-14' '1.03625144071e-13' '-5.52444866686e-14' '-5.04415946628e-14' '1.9026074286e-14' '3.42584334296e-14' 'apple']

Unable to comprehend why it is not being saved.

I already looked at : numpy beginner: writing an array using numpy.savetxt numpy savetxt fails when using file handler How to combine a numpy array and a text column and export to csv

Please guide me how to save this new numpy array properly. I'm from an R programming background, in python is there any easy of python equivalent of saving this array like an R data frame kind of structure?

Final goal is to send this into a neural network.

I tried specifying 'fmt = %s' as an option but nothing happens. That's awfully vague. No error message? No file output? Are you sure you are looking in the right place for that file? — hpaulj
– hpaulj, Commented Feb 9, 2018 at 4:15
I too found that weird , there is no error no warning absolutely nothing. I even tried ` locate mfcc_with_labels` ` from terminal . nothing showed up. Let me restart my laptop and try again/. — kRazzy R
– kRazzy R, Commented Feb 9, 2018 at 4:18

hpaulj · Accepted Answer · 2018-02-09 04:10:44Z

2

The default fmt for savetxt is %.18e. Try that with a number

In [84]: '%.18e'%12
Out[84]: '1.200000000000000000e+01'

The actual format is that string replicated 21 times (the number of columns) and joined with the delimiter.

But your array has a string dtype, and contains strings (because you appended the labels. That doesn't work with that format.

Your mfcc_with_labels[1]

In [86]:     row = np.array(['-5.04415946628e-14', '1.9026074286e-14', '3.425843
    ...: 34296e-14', 'apple'])
In [87]: row
Out[87]: 
array(['-5.04415946628e-14', '1.9026074286e-14', '3.42584334296e-14',
       'apple'], dtype='<U18')

'%s' fmt should work; this formatting does:

In [88]: '%s,%s,%s,%s'%tuple(row)
Out[88]: '-5.04415946628e-14,1.9026074286e-14,3.42584334296e-14,apple'

answered Feb 9, 2018 at 4:10

hpaulj

233k14 gold badges260 silver badges392 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

kRazzy R Over a year ago

I see the point you are trying to make. the fmt =%s worked. For some reason ubuntu was not able to locate the files. a restart solved the issue.

Collectives™ on Stack Overflow

unable to do np.savetxt on save newly appended string column to numpy array

1 Answer 1

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related