profile_python_driver.py

import cProfile, pstats, StringIO
from itertools import cycle

from cassandra.concurrent import execute_concurrent
from cassandra.cluster import Cluster


concurrency = 50
TABLE = "testtable"
KEYSPACE = "testkeyspace"
query = "INSERT INTO {0}(k, v) VALUES (%s, %s)".format(TABLE)

warmup_num_ops = int(1e4)
num_ops = int(1e5)

statements = cycle((query,))

warmup_parameters = [(i, i + 1) for i in range(warmup_num_ops)]
warmup_query_and_params = list(zip(statements, warmup_parameters))


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

drop_keyspace ='''DROP KEYSPACE IF EXISTS %s''' % KEYSPACE
session.execute(drop_keyspace)

create_keyspace = '''
   CREATE KEYSPACE %s
   WITH replication = {'class': 'SimpleStrategy', 'replication_factor': '3'}''' % KEYSPACE
session.execute(create_keyspace)

session.set_keyspace(KEYSPACE)

create_table = '''
            CREATE TABLE %s.%s (
                k int PRIMARY KEY,
                v int )''' % (KEYSPACE, TABLE)
session.execute(create_table)

# warmup
print("Starting warming")
execute_concurrent(session, warmup_query_and_params,
               concurrency=concurrency, results_generator=False)

parameters = [(i, i + 1) for i in range(num_ops)]
query_and_params = list(zip(statements, parameters))

print("Starting profiling")
pr = cProfile.Profile()
pr.enable()

for _ in range(100):
    execute_concurrent(session, query_and_params, concurrency=concurrency, results_generator=False)

pr.disable()
s = StringIO.StringIO()
sortby = 'cumulative'
f = open("stats_execute_concurrent.txt", "w")
ps = pstats.Stats(pr, stream=f).sort_stats(sortby)
ps.print_stats()
f.close()