Making a numpy array with mixed types and writing it as csv

Question

I am trying to create a numpy array to hold different types like this:

na_csv_output = np.zeros((len(ldt_timestamps),1),dtype=('i4,i4,i4,a10,a10,i4'))

The problem with this is that all 6 values are stored as a single entry, whereas I would prefer it to be 6 columns so that they can be written properly as a CSV. Is there any way this can be done by using a numpy array? (or some other type like list)

P.S.

Even when I try to write the values that I get from the initial try to a file using

np.savetxt('eventResults.csv', na_csv_output, delimiter=",")

it says

TypeError: float argument required, not numpy.string_

so that's another problem.

gg349 · Accepted Answer · 2013-09-30 23:09:20Z

2

To access each column, use na_csv_output['f0'] or f1 f2, etc, where f stands for field.

To save it to a file, first make it a 1D array,

na_csv_output = np.zeros((len(ldt_timestamps)),dtype=('i4,i4,i4,a10,a10,i4'))

To save it, you can go for something like this for example

np.savetxt('eventResults.csv', na_csv_output, fmt='%f, %f, %f, string= %s %s, last number = %f', delimiter=",")

answered Sep 30, 2013 at 23:09

gg349

22.8k5 gold badges58 silver badges65 bronze badges

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

2 Comments

mzm Over a year ago

That's exactly what I need, but it's not working correctly, if I try to make it 1D and insert elements like na_csv_output[0] = [(1,1,1,'a','b',1)] it says TypeError: expected a readable buffer object, and if I make it na_csv_output = np.zeros((len(ldt_timestamps),1),dtype=('i4,i4,i4,a10,a10,i4')) it says AttributeError: fmt has wrong number of % formats. %f, %f, %f, %s, %s, %f

gg349 Over a year ago

Use instead na_csv_output[0]=(1,1,1,'a','b',1) for an assignment, and stick with a 1D array.

mattexx · Accepted Answer · 2013-10-01 00:06:41Z

0

I would use pandas. It makes csv operations really easy.

import pandas as pd
# omit second argument to np.zeros to get 1-d data
na_csv_output = np.zeros((len(ldt_timestamps),),dtype=('i4,i4,i4,a10,a10,i4'))
df = pd.DataFrame(na_csv_output)
df.to_csv('eventResults.csv')

answered Oct 1, 2013 at 0:06

mattexx

6,6163 gold badges40 silver badges47 bronze badges

Comments

hpaulj · Accepted Answer · 2013-10-01 03:29:52Z

0

with open('eventResults.csv','w') as f:
    for x in xx:
        np.savetxt(f,x,fmt='%f, %f, %f, %r, %r, %f')
        # np.savetxt(f,x,fmt='%r',delimiter=', ')

edited Oct 1, 2013 at 3:29

answered Oct 1, 2013 at 3:23

hpaulj

233k14 gold badges260 silver badges392 bronze badges

Collectives™ on Stack Overflow

Making a numpy array with mixed types and writing it as csv

3 Answers 3

2 Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

2 Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related