Pulling out data from CSV files' specific columns in Python

Question

I need a quick help with reading CSV files using Python and storing it in a 'data-type' file to use the data to graph after storing all the data in different files.

I have searched it, but in all cases I found, there was headers in the data. My data does not header part. They are tab separated. And I need to store only specific columns of the data. Ex:

12345601 2345678@abcdef 1 2 365 places

In this case, as an example, I would want to store only "2345678@abcdef" and "365" in the new python file in order to use it in the future to create a graph.

Also, I have more than 1 csv file in a folder and I need to do it in each of them. The sources I found did not talk about it and only referred to:

# open csv file
with open(csv_file, 'rb') as csvfile:

Could anyone refer me to already answered question or help me out with it?

Answer 1

. . . and storing it in a PY file to use the data to graph after storing all the data in different files . . .

. . . I would want to store only "2345678@abcdef" and "365" in the new python file . . .

Are you sure that you want to store the data in a python file? Python files are supposed to hold python code and they should be executable by the python interpreter. It would be a better idea to store your data in a data-type file (say, preprocessed_data.csv ).

To get a list of files matching a pattern, you can use python's built-in glob library .

Here's an example of how you could read multiple csv files in a directory and extract the desired columns from each one:

import glob

# indices of columns you want to preserve
desired_columns = [1, 4]
# change this to the directory that holds your data files
csv_directory = '/path/to/csv/files/*.csv'

# iterate over files holding data
extracted_data = []
for file_name in glob.glob(csv_directory):
    with open(file_name, 'r') as data_file:
        while True:
            line = data_file.readline()
            # stop at the end of the file
            if len(line) == 0:
                break

            # splits the line by whitespace
            tokens = line.split()
            # only grab the columns we care about
            desired_data = [tokens[i] for i in desired_columns]
            extracted_data.append(desired_data)

It would be easy to write the extracted data to a new file. The following example shows how you might save the data to a csv file.

output_string = ''
for row in extracted_data:
    output_string += ','.join(row) + '\n'

with open('./preprocessed_data.csv', 'w') as csv_file:
    csv_file.write(output_string)

Edit:

If you don't want to combine all the csv files, here's a version that can process one at a time:

def process_file(input_path, output_path, selected_columns):
    extracted_data = []    
    with open(input_path, 'r') as in_file:
        while True:
            line = in_file.readline()
            if len(line) == 0: break
            tokens = line.split()
            extracted_data.append([tokens[i] for i in selected_columns])

    output_string = ''
    for row in extracted_data:
        output_string += ','.join(row) + '\n'

    with open(output_path, 'w') as out_file:
        out_file.write(output_string)

# whenever you need to process a file:
process_file(
    '/path/to/input.csv', 
    '/path/to/processed/output.csv',
    [1, 4])

# if you want to process every file in a directory:
target_directory = '/path/to/my/files/*.csv'
for file in glob.glob(target_directory):
    process_file(file, file + '.out', [1, 4])

Edit 2:

The following example will process every file in a directory and write the results to a similarly-named output file in another directory:

import os
import glob

input_directory = '/path/to/my/files/*.csv'
output_directory = '/path/to/output'
for file in glob.glob(input_directory):
    file_name = os.path.basename(file) + '.out'
    out_file = os.path.join(output_directory, file_name)
    process_file(file, out_file, [1, 4])

If you want to add headers to the output, then process_file could be modified like this:

def process_file(input_path, output_path, selected_columns, column_headers=[]):
    extracted_data = []    
    with open(input_path, 'r') as in_file:
        while True:
            line = in_file.readline()
            if len(line) == 0: break
            tokens = line.split()
            extracted_data.append([tokens[i] for i in selected_columns])

    output_string = ','.join(column_headers) + '\n'
    for row in extracted_data:
        output_string += ','.join(row) + '\n'

    with open(output_path, 'w') as out_file:
        out_file.write(output_string)

Answer 2

Here's another approach using a namedtuple that will help extract selected fields from a csv file and then let you write them out to a new csv file.

from collections import namedtuple    
import csv

# Setup named tuple to receive csv data
# p1 to p5 are arbitrary field names associated with the csv file
SomeData = namedtuple('SomeData', 'p1, p2, p3, p4, p5, p6')

# Read data from the csv file and create a generator object to hold a reference to the data
# We use a generator object rather than a list to reduce the amount of memory our program will use
# The captured data will only have data from the 2nd & 5th column from the csv file
datagen = ((d.p2, d.p5) for d in map(SomeData._make, csv.reader(open("mydata.csv", "r"))))

# Write the data to a new csv file
with open("newdata.csv","w", newline='') as csvfile:
    cvswriter = csv.writer(csvfile, delimiter=',', quotechar='"', quoting=csv.QUOTE_MINIMAL)

    # Use the generator created earlier to access the filtered data and write it out to a new csv file
    for d in datagen:
        cvswriter.writerow(d)

Original Data in "mydata.csv":

12345601,2345678@abcdef,1,2,365,places  
4567,876@def,0,5,200,noplaces

Output Data in "newdata.csv":

2345678@abcdef,365  
876@def,200

EDIT 1: For tab delimited data make the following changes to the code:
change
datagen = ((d.p2, d.p5) for d in map(SomeData._make, csv.reader(open("mydata.csv", "r"))))
to
datagen = ((d.p2, d.p5) for d in map(SomeData._make, csv.reader(open("mydata2.csv", "r"), delimiter='\\t', quotechar='"')))
and
cvswriter = csv.writer(csvfile, delimiter=',', quotechar='"', quoting=csv.QUOTE_MINIMAL)
to
cvswriter = csv.writer(csvfile, delimiter='\\t', quotechar='"', quoting=csv.QUOTE_MINIMAL)