Riak's "Loading Data and Running MapReduce Queries" for Python

load_data.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

# see http://wiki.basho.com/Loading-Data-and-Running-MapReduce-Queries.html

import sys
import riak

HOST = '127.0.0.1'
PORT = '8098'

headers = [u'Date', u'Open', u'High', u'Low', u'Close', u'Volume', u'Close']

def main(filename):
    cl = riak.RiakClient(host=HOST, port=PORT)
    bkt = cl.bucket('goog')
    
    lines = read_file(filename)
    for line in lines[1:]:
        key, data = format_data(line)
        bkt.new(key, data=data).store()
        print('inserted: {0}'.format(key))

def read_file(filename):
    lines = None
    with open(filename, 'r') as f:
        lines = [line for line in f]
    return lines

def format_data(line):
    data = line.split(',')
    json = {}
    for i, header in enumerate(headers):
        json[header] = data[i].strip()
    return json[u'Date'], json

if __name__ == '__main__':
    if len(sys.argv) != 2:
        print('usage: python load_data.py <filename>')
        sys.exit(0)

    filename = sys.argv[1]
    main(filename)