CODEWITHSUNDEEP
pythonarraysnumpypandas
avicohen
avicohen

Reputation: 3047

How to convert a numpy matrix to a pandas series?

I have the following numpy matrix:

array([[64, 22,],
   [58, 64],
   [42, 31])

And i want to get the following:

pd.DataFrame({'one':"64 22", 'two':"42 31"})

My purpose is to convert each row in the numpy.array to a string that will be used for a pandas dataframe. Is there some built in pandas function to the rescue?

Upvotes: 5

Views: 20022

Answers (1)

jezrael
jezrael

Reputation: 862511

IIUC you can use DataFrame constructor and apply join:

import pandas as pd
import numpy as np

arr = np.array([[64, 22,],   [58, 64],   [42, 31]])
print arr
[[64 22]
 [58 64]
 [42 31]]

li = ['one','two','three']
df = pd.DataFrame(arr, dtype='str', index=li)
print df
        0   1
one    64  22
two    58  64
three  42  31

print df.apply(lambda x: ' '.join(x), axis=1)
one      64 22
two      58 64
three    42 31
dtype: object

Upvotes: 5

Related Questions