在python中获取没有numpy或pandas的数组的样本

Question

I am trying to create a "pythonic" way of taking a small subset of a very large array in python.

I am currently taking in a csv with 58 columns and 4960 rows with the following codes:

def import_normal_csv(file):
    # Create blank array
    results = []
    # Open file
    with open(file) as csvfile:
        # read in file changing values to floats
        reader = csv.reader(csvfile, quoting=csv.QUOTE_NONNUMERIC)
        for row in reader:
            results.append(row)
    return results

def main():
    print(" Working SPAM Dataset... ")
    # Create a raw data array without numpy
    spam_raw_data = import_normal_csv('spam.csv')

    # CREATE SUBSET OF SPAM_RAW_DATA HERE

    random.shuffle(spam_raw_data)

I have seen various ways to do this using numpy or pandas , but I would like to do this naturally without those libraries. Instead of my massive array, how could I instead take in only...500 rows (or some arbitrary number significantly less than nearly 5000)?

Answer 1

You can use the builtin random library, for example:

import random

random.sample(data, 500)

This will give you a list of 500 list s, each representing one row.

Answer 2

Use random.sample :

subset_size = 500
random.sample(spam_raw_data, subset_size)

Also note your import_normal_csv function can be simplified:

def import_normal_csv(file):
    with open(file) as csvfile:
        reader = csv.reader(csvfile, quoting=csv.QUOTE_NONNUMERIC)
        return list(reader)