Insertion script

insert_smallsavings.py

"""Simple Data loader.

Usage:
  insert_smallsavings.py [-c] [-n=<x>]

Options:
  -c  enable compression (disabled by default)
  -n=<x> total number of rows to insert (default: 1M)
"""
from __future__ import print_function
from uuid import uuid4, uuid1
from docopt import docopt
from cassandra.cluster import Cluster
from random import randint, choice, random
from string import ascii_letters

max_async_query = 1000

if __name__ == '__main__':
    args = docopt(__doc__)
    use_compression = args['-c']

    rows = 1000000 if args['-n'] is None else args['-n']

    cluster = Cluster()
    session = cluster.connect()

    session.execute("CREATE KEYSPACE ks WITH replication = { 'class': 'SimpleStrategy', 'replication_factor': '1' };")

    session.execute("""
        CREATE TABLE ks.smallsavings (
            k int PRIMARY KEY,
            v1 int,
            v2 text
        ) WITH compression = {'sstable_compression': '%s'}
        """ % ('LZ4Compressor' if use_compression else ''))

    insert = session.prepare('INSERT INTO ks.smallsavings (k, v1, v2) VALUES (?, ?, ?)')

    futures = []
    for row in range(0, rows):
        random_text = ''.join(choice(ascii_letters) for _ in range(1000))
        futures.append(session.execute_async(insert, [row, randint(0, 2^31), random_text]))
        if (len(futures) > max_async_query):
            for f in futures:
                f.result()
            futures = []

    for f in futures:
        f.result()

    assert len(list(session.execute('SELECT * from ks.smallsavings;'))) == rows