Convert elements in a numpy array to string

Question

I wrote the following script to load data from CSV file in numpy array form:

import numpy
sample = numpy.genfromtxt('sample_cor.csv', delimiter = ',')
print sample

and this sample looked like:

[[ 259463.392 2737830.062 ]
 [ 255791.4823 2742050.772 ]
 [ 249552.4949 2746152.328 ]
 [ 247925.1228 2746422.143 ]
 [ 262030.4697 2728966.229 ]
 [ 260462.1936 2731412.856 ]
 [ 260644.0281 2735003.027 ]
 [ 268588.7974 2732835.097 ]]

now I want to extract every row in this array to string with a comma, for example, I expected row 1 to be converted to 259463.392,2737830.062, row 2 to be 255791.4823,2742050.772, and so on.

I tried the code below:

ss = numpy.array_str(sample[0])
print ss
print type(ss)

and got the result maybe not what I want,

[ 259463.392 2737830.062]
<type 'str'>

(I used coords = '259467.2,2737833.87' and got the string form which was what I want: 259467.2,2737833.87)

How to convert elements in a numpy array to string with a comma?

• hpaulj: The form without the , is just the normal way that numpy arrays are displayed. It helps distinguish them from Python lists. The , is not an integral part of either a list or an array. It's just a display feature.

Answer 1

Here's an approach using join method -

[",".join(item) for item in a.astype(str)]

Sample run -

In [141]: a
Out[141]: 
array([[ 259463.392 , 2737830.062 ],
 [ 255791.4823, 2742050.772 ],
 [ 249552.4949, 2746152.328 ],
 [ 247925.1228, 2746422.143 ],
 [ 262030.4697, 2728966.229 ],
 [ 260462.1936, 2731412.856 ],
 [ 260644.0281, 2735003.027 ],
 [ 268588.7974, 2732835.097 ]])
In [142]: [",".join(item) for item in a.astype(str)]
Out[142]: 
['259463.392,2737830.062',
 '255791.4823,2742050.772',
 '249552.4949,2746152.328',
 '247925.1228,2746422.143',
 '262030.4697,2728966.229',
 '260462.1936,2731412.856',
 '260644.0281,2735003.027',
 '268588.7974,2732835.097']

• Heinz: thank you for the advice, but I have a little question that why the showing of mine of numpy array in the post lack comma?
• Divakar: @Heinz Hmm if the input is a NumPy array, it should. Or else, you could try print(repr(a)).
• Heinz: Thank you, repr() returns the having-comma array.
• fazistinho_: @Divakar is there a "pure" numpy solution to this and would that make it faster for large datasets? Thank you
• Rishabh Barman: I am very late to the party, though! hey @Heinz, since you're printing that Numpy array via print(), it doesn't display the commas. Try simply running the variable in a single Jupyter cell - then, elements display with commas.