Define a train/test threshold:

ex: 80% train - 20% test



In [6]:

    
split = .8

Specify your input file:



In [7]:

    
file_origin = "/path/to/your/input/dataset.csv"

Specify your ouput files:



In [ ]:

    
file_train = "/path/to/your/output/train_dataset.csv"
file_test = "/path/to/your/output/test_dataset.csv"

Execute the following python code:



In [3]:

    
from random import random



In [8]:

    
with open(file_train, 'w') as train,\
     open(file_test, 'w') as test,\
     open(file_origin) as origin:
    for line in origin:
        rand = random()
        if rand < split:
            train.write(line)
        else:
            test.write(line)