Flatten additional_attributes from Magento 2 export csv

flatten_additional_attributes.py

"""
This script formats a magento product export file, and sets each
additional_attribute as a new column in a target file
"""
import csv

# real files
source_file = './catalog_product_20210316_044037.csv'
target_file = './catalog_product_20210316_044037_formatted.csv'

def get_modified_headers():
    reader = csv.reader(open(source_file, 'r'))
    headers = next(reader)

    # get all headers for additional_attributes
    additional_attributes_headers = []
    i = 0
    for row in reader:
        target_index = headers.index("additional_attributes")

        attributes = row[target_index].split(",")
        for attribute in attributes:
            data = attribute.split("=")
            if len(data) != 1:
                if data[0] != 'publish_date':
                    header = data[0]
                    # add attribute as header
                    if header not in additional_attributes_headers:
                        additional_attributes_headers.append(header)

        # iterator
        i += 1

    headers_modified = headers + additional_attributes_headers

    return [headers_modified, additional_attributes_headers]


def run_with_writer():
    """ Main run function """
    reader = csv.reader(open(source_file, 'r'))
    writer = csv.writer(open(target_file, 'w'))

    # get original headers
    headers = next(reader)

    # add additional attributes to headers
    modified_headers = get_modified_headers()[0]
    additional_attributes_headers = get_modified_headers()[1]

    # write headers to file
    writer.writerow(modified_headers)

    # extract data and populate rows
    i = 0
    for row in reader:
        print(f'processing row: {i}')

        result = [None] * len(additional_attributes_headers)
        # pre-populate result (to avoid shifting items as we add insert them at the index)
        target_index = headers.index("additional_attributes")
        attributes = row[target_index].split(",")

        for attribute in attributes:
            data = attribute.split("=")

            if len(data) != 1:
                if data[0] != 'publish_date':
                    header = data[0]
                    value = data[1]

                    target_index = additional_attributes_headers.index(header)
                    result[target_index] = value


        modified_row = row + result
        
        # write modified row
        writer.writerow(modified_row)

        i += 1


# Main entry point
if __name__ == '__main__':
    run_with_writer()