use np.savetxt to save structured numpy array containing string and float

Question

I have a numpy structured array which contains string and float. I want to save this structured array as it is into a csv file. The simplified version of my procedure is like this.

    structured_array = np.zeros((1,), dtype=[('string','a20'),('float','f8')])
    structured_array['string'] = 'string'
    structured_array['float'] = 0.0
    np.savetxt('foo.csv', structured_array, delimiter=',',fmt='%s,%f')

I would expect string,0.000000 in foo.csv, but it gives me b'string',0.000000 where does this quotation mark and this b comes from? How can I get rid of it?

I can use readline() and manually get rid of this but is there any clever way to do this.

Thank you very much.

I can't replicate... I get string,0.000000 have you got some sort of encoding set to your script? try ('string','S20') for the first field — user1121588
– user1121588, Commented Jul 10, 2015 at 16:04
@DanPatterson I can replicate, numpy 1.9.2, and changing to 'S20' has no effect. — Matthew Plourde
– Matthew Plourde, Commented Jul 10, 2015 at 17:35
@MatthewPlourde my version is 1.7.x could this be a Unicode difference issue? — user1121588
– user1121588, Commented Jul 10, 2015 at 19:46

Matthew Plourde · Accepted Answer · 2015-07-10 17:33:11Z

1

Line 1087 in savetxt (...\lib\site-packages\numpy\lib\npio.py) has

for row in X:
    fh.write(asbytes(format % tuple(row) + newline))

Which reveals that columns are converted to bytes before writing (hence the b prefix. It doesn't appear that this can be changed.

answered Jul 10, 2015 at 17:33

Matthew Plourde

44.8k9 gold badges103 silver badges116 bronze badges

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

Collectives™ on Stack Overflow

use np.savetxt to save structured numpy array containing string and float

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related