How to convert vector wrapped as string to numpy array in pandas dataframe?

Question

I have a pandas dataframe with a column of vectors that I would like to perform matrix arithmetic on. However, upon closer inspection the vectors are all wrapped as strings with new line characters seemingly embedded in them:

How do I convert each vector in this column into numpy arrays? I've tried

df['Word Vector'].as_matrix

and

np.array(df['Word Vector'])

as well as

df['Word Vector'] = df['Word Vector'].astype(np.array)

but none produced the desired result. Any pointers would be appreciated!

Answer 1

Hope the following works as what you expected

import pandas as pd
import numpy as np

x = str(np.arange(1,100))
df = pd.DataFrame([x,x,x,x])
df.columns = ['words']
print 'sample'
print df.head()
result = df['words'].apply(lambda x: 
                           np.fromstring(
                               x.replace('\n','')
                                .replace('[','')
                                .replace(']','')
                                .replace('  ',' '), sep=' '))
print 'result'
print result

output as following

    sample
                                               words
0  [ 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 ...
1  [ 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 ...
2  [ 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 ...
3  [ 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 ...
result
0    [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, ...
1    [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, ...
2    [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, ...
3    [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, ...

It is not elegant to call replace function so many times. However I did not find better approach. Anyway it should help you to convert string to vectors.

A side note, as data is presented in picture, You'd better check whether your data separation is done by space or tab. If it is tab, change sep=' ' to sep='\t'

Answer 2

This worked for me for string lists in a Pandas column:

df['Numpy Word Vector'] = df['Word Vector'].apply(eval).apply(np.array)