intersection_test.py

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import time

data = []
for n_elements in [100, 1000, 10000, 100000]:
    a = range(n_elements)
    sa = set(a)
    b = np.array(a) + 50
    sb = set(b.tolist())
    start = time.time()
    c = []
    for e in a:
        if e in b:
            c.append(e)
    stop = time.time()
    list_timing = stop - start
    print "{0} Elements intersected in {1} seconds.".format(n_elements, list_timing)
    print "Intersection result: {0}, last three elements: {1}.".format(len(c), c[-3:])
    
    # Set intersection
    start = time.time()
    c2 = sa.intersection(sb)
    stop = time.time()
    set_timing = stop - start
    print "Set Intersect - {0} Elements intersected in {1} seconds.".format(n_elements, set_timing)
    print "Intersection result: {0}, last three elements: {1}.".format(len(c2), list(c2)[-3:])
    print "-"*70
    data.append({"elements" : n_elements, "list_timing" : list_timing, "set_timing" : set_timing})

#plotting it for convenience
df = pd.DataFrame(data)
df = df.set_index("elements")
plt.style.use("ggplot")
df.plot(style="o-", logx=True)
plt.ylabel("seconds")
plt.title("Quick benchmark test")
plt.show()