How to copy unique individuals from a csv file into new files

Question

I am very new to python so please bear with me. I have a csv file that looks like this:

.

I am trying to loop through the file and for every unique individual there are, create a new csv file and copy rows over. I was successful in doing this for one animal but I am having trouble creating the syntax for a more generic approach. Here is what I currently have:

import arcpy
import csv
from csv import DictReader

WS = arcpy.env.workspace = raw_input("Where if your workspace")
infile = raw_input("where is your file?") 
outfile = raw_input("What is your outfile name?") 
arcpy.env.overwriteOutput = True


with open(infile, "r") as csvFile, open(outfile, "w") as out, open("outfile2.csv", "w") as out2:
    reader = csv.DictReader(csvFile)
    writer = csv.writer(out)
    writer.writerow(reader.fieldnames)
    for row in reader:
         if row["Animal"] == "1":
            values = [row[field] for field in reader.fieldnames]
            writer.writerow(values)

Answer 1

To write each Animal into its own CSV file, you would need to open a different file for each type of animal. This could be done by using a dictionary to store the file object and csv writer object for each animal. At the end, this could then be used to close all the files correctly:

import csv

output_csvs = {}    # e.g. {'1' : [file_object, csv_object]}

with open('input.csv', 'rb') as f_input:
    csv_reader = csv.reader(f_input)
    header = next(csv_reader)

    for row in csv_reader:
        animal = row[0]

        if animal in output_csvs:
            output_csvs[animal][1].writerow(row)
        else:
            f_output = open('animal_{}.csv'.format(animal), 'wb')
            csv_output = csv.writer(f_output)
            output_csvs[animal] = [f_output, csv_output]
            csv_output.writerow(header)
            csv_output.writerow(row)

for csv_file, csv_writer in output_csvs.values():
    csv_file.close()

This would give you a set of output CSV files named according to the animal, e.g. animal_1.csv

---

Alternatively, if the data is small enough to be read into memory, it could be sorted by animal and output one block at a time by making use of Python's itertools.groupby() function:

from itertools import groupby
import csv

with open('input.csv', 'rb') as f_input:
    csv_reader = csv.reader(f_input)
    header = next(csv_reader)

    for animal, group in groupby(sorted(csv_reader), lambda x: x[0]):
        with open('animal_{}.csv'.format(animal), 'wb') as f_output:
            csv_output = csv.writer(f_output)
            csv_output.writerow(header)
            csv_output.writerows(group)

Using sorted() ensures that all animals of the same kind are grouped together. If this is already the case in the data, sorting is not needed.

---

To access these files you could use glob.glob():

import matplotlib.pyplot as plt            
import glob

for animal_filename in glob.glob('animal_*.csv'):
    with open(animal_filename, 'rb') as f_input:
        csv_input = csv.reader(f_input)
        heading = next(csv_input)
        x, y = [], []

        for row in csv_input:
            x.append(int(row[1]))
            y.append(int(row[2]))

        fig, ax = plt.subplots()
        plt.title(animal_filename)
        ax.scatter(x, y)

plt.show()