comparing different string search methods

results.txt

----------------------------------------------------------------------------------- benchmark: 3 tests -----------------------------------------------------------------------------------
Name (time in us)                  Min                   Max                  Mean              StdDev                Median                IQR            Outliers(*)  Rounds  Iterations
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_search[re_compile]       452.1870 (1.0)      1,199.2530 (1.13)       471.1449 (1.0)       56.5338 (1.0)        455.0175 (1.0)       3.9840 (1.0)          120;321    1844    1
test_search[no_compile]       521.2920 (1.15)     1,059.7880 (1.0)        554.4658 (1.18)      92.9603 (1.64)       526.5430 (1.16)      5.2185 (1.31)           28;75     388    1
test_search[endswith]       1,601.1760 (3.54)     3,252.0090 (3.07)     1,691.4124 (3.59)     246.1582 (4.35)     1,604.0630 (3.53)     38.6730 (9.71)          39;102     520    1
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

test_search.py

import random as rnd
from itertools import cycle, product
import re

import pytest


def search_re_no_compile(search_items, strings_to_search):
    p = '(' + '|'.join(search_items) + ')$'
    return tuple(re.search(p, s) for s in strings_to_search)


def search_re_compile(search_items, strings_to_search):
    p = re.compile('(' + '|'.join(search_items) + ')$')
    return tuple(p.search(s) for s in strings_to_search)


def search_simple(search_items, strings_to_search):
    return tuple(
        next((True for si in search_items if s.endswith(si)), False)
        for s in strings_to_search
    )


search_items = tuple(str(1000 + i) for i in range(100))
random_data = tuple(str(rnd.randint(0, 100000)) * 100 for i in range(10000))
strings_to_search = tuple(
    si + rd if should_match else si
    for si, rd, should_match in
    zip(search_items, random_data, cycle((True, False)))
)


@pytest.mark.parametrize(
    'search_func',
    (search_re_no_compile, search_re_compile, search_simple),
    ids=('no_compile', 're_compile', 'endswith')
)
def test_search(benchmark, search_func):
    benchmark(search_func, search_items, strings_to_search)