Sebastian Raschka
last updated: 05/29/2014

I would be happy to hear your comments and suggestions.
Please feel free to drop me a note via twitter, email, or google+.

Day 9 - One Python Benchmark per Day

The most Pythonic way to check if a string ends with a particular substring

-> skip to the results

This benchmark was inspired by the article "A Nice Little Bit of Python" for finding the most elegant way to determine if a string ends with a particular substring.
There is the "old-fashioned" way using a sequence of "or"s, e.g.,

if needle.endswith('ly') or needle.endswith('ed') or\
    needle.endswith('ing') or needle.endswith('ers'):
    print('Is valid')
else:
    print('Invalid')

Or the more elegant way using a list comprehension that was suggested by the author:

if any([needle.endswith(e) for e in ('ly', 'ed', 'ing', 'ers')]):
    print('Is valid')
else:
    print('Invalid')

However, there are two even better solutions. One would be replacing the list comprehension by a generator, and my favorite one: feeding a tuple to the .endswith() method. See the different implementations below:

Functions for checking if a string ends with a substring



In [1]:

    
from re import search as re_search
from re import compile as re_compile

def old_fashioned(needle):
    return bool(needle.endswith('ly') or needle.endswith('ed') or\
        needle.endswith('ing') or needle.endswith('ers'))

def list_comprehension(needle):
    return bool(any([needle.endswith(e) for e in ('ly', 'ed', 'ing', 'ers')]))
        
def generator(needle):
    return bool(any(needle.endswith(e) for e in ('ly', 'ed', 'ing', 'ers')))

def endswith_tuple(needle):
    return bool(needle.endswith(('ly', 'ed', 'ing', 'ers')))

def regexpr(needle):
     return bool(re_search(r'(?:ly|ed|ing|ers)$', needle))

comp = re_compile(r'(?:ly|ed|ing|ers)$')    
def compiled_regexpr(needle):
    return bool(comp.search(needle))
    
def map_func(needle):
    return any(map(needle.endswith, ('ly', 'ed', 'ing', 'ers')))

Quick note on why I am importing re.search as re_search:
To decrease the overhead for the lookup.



In [2]:

    
import re

%timeit re_search(r'(?:ly|ed|ing|ers)$', 'needlers')
%timeit re.search(r'(?:ly|ed|ing|ers)$', 'needlers')









    



1000000 loops, best of 3: 1.63 µs per loop
1000000 loops, best of 3: 1.69 µs per loop

Verification that all functions work correcltly



In [3]:

    
funcs = [old_fashioned, list_comprehension, 
         generator, endswith_tuple, 
         regexpr, compiled_regexpr,
         map_func
        ]



In [4]:

    
for f in funcs:
    assert(f('neeeeedly') == True), f.__name__
    assert(f('neeeeedlee') == False), f.__name__
    assert(f('nelyeedlee') == False), f.__name__
print('All functions work correctly.')









    



All functions work correctly.

Timing



In [5]:

    
import timeit

test_cases = ["neeeeedly", "neeeeedlers", "neeeeedlee", ]
times_n = {f.__name__:[] for f in funcs}

for t in test_cases:
    for f in funcs:
        f = f.__name__
        times_n[f].append(min(timeit.Timer('%s(t)' %f, 
                      'from __main__ import %s, t' %f)
                              .repeat(repeat=500, number=1000)))



In [6]:

    
import platform
import multiprocessing

def print_sysinfo():
    
    print('\nPython version  :', platform.python_version())
    print('compiler        :', platform.python_compiler())

    print('\nsystem          :', platform.system())
    print('release         :', platform.release())
    print('machine         :', platform.machine())
    print('processor       :', platform.processor())
    print('CPU count       :', multiprocessing.cpu_count())
    print('interpreter     :', platform.architecture()[0])
    print('\n\n')



In [7]:

    
%matplotlib inline



In [10]:

    
from numpy import arange
import matplotlib.pyplot as plt

def plot():

    labels = [('old_fashioned','needle.endswith("ly") or needle.endswith("ed") or ...'),
          ('list_comprehension', 'any([needle.endswith(e) for e in ("ly", "ed", ...)])'),
          ('generator', 'any(needle.endswith(e) for e in ("ly", "ed", ...))'),
          ('regexpr', 're.search(r"(?:ly|ed|ing|ers)$", needle)'),
          ('compiled_regexpr', 'same as above, but with compiled regexpr'),
          ('map_func', 'any(map(needle.endswith, ("ly", "ed", "ing", "ers")))'),
          ('endswith_tuple', 'needle.endswith(("ly", "ed", "ing", "ers"))'),
          
          ]

    x_labels = ['True (first match): "neeeeedly"',
                'True (first match): "neeeeedlers"',
                'False: "neeeeedlee"'
                ]
            

    ind = arange(len(test_cases))  # the x locations for the groups
    width = 0.12

    fig = plt.figure(figsize=(14,8))
    ax = fig.add_subplot(111)
    colors = [(0,'c'), (1,'b'), (2,'g'), (3,'r'), (4,'y'), (5,'m'), (6, 'k')]
    
    for l,c in zip(labels,colors):
        ax.bar(ind + c[0]*width,
            times_n[l[0]], 
            width,
            alpha=0.5,
            color=c[1],
            label=l[1]
        )

    ax.set_ylabel('execution time in microseconds', fontsize=16)
    ax.set_title('Methods for determening if a string '\
                 'ends with a particular substring',
                 fontsize=18)
    ax.set_xticks(ind + width + 0.24)
    ax.set_xticklabels(x_labels)
    leg = ax.legend(loc='upper left', bbox_to_anchor = (0.85, 1.0))
    plt.grid()
    plt.ylim([0, 0.003])
    plt.xlim([-0.2, 3.0])
    plt.show()

Results



In [11]:

    
print_sysinfo()
plot()









    



Python version  : 3.3.5
compiler        : GCC 4.0.1 (Apple Inc. build 5493)

system          : Darwin
release         : 13.2.0
machine         : x86_64
processor       : i386
CPU count       : 4
interpreter     : 64bit

Conclusion

Clearly, the most elegant and Pythonic way of determining if a string ends with a particular substring is also the most efficient one!