LEARN COMPLETE PYTHON IN 24 HOURS

12. Performance Optimization

12.1 Time & Space Complexity Basics

Time Complexity — How runtime grows with input size (n) Space Complexity — How memory usage grows with input size

Common Big-O notations (from best to worst):

NotationNameGrowth Rate Example (n = 10 → 1,000)When you see itO(1)ConstantSame time alwaysDictionary lookup, array accessO(log n)LogarithmicVery slow growthBinary search, balanced tree opsO(n)LinearDoubles when input doublesLooping once over listO(n log n)LinearithmicFast for large nEfficient sorting (TimSort)O(n²)Quadratic100× slower when n×10Nested loops (bubble sort, etc.)O(2ⁿ)ExponentialExplodes very fastRecursive Fibonacci (naive)

Quick examples:

Python

# O(1) – constant time def get_user(users, user_id): return users.get(user_id) # dict lookup # O(n) – linear time def find_max(lst): return max(lst) # loops once internally # O(n²) – quadratic (avoid for large n) def has_duplicates(lst): for i in range(len(lst)): for j in range(i+1, len(lst)): if lst[i] == lst[j]: return True return False # O(n log n) – good sorted_list = sorted(my_list) # uses TimSort

Rule of thumb (2026):

• n ≤ 10³ → almost anything is fine

• n ≈ 10⁵–10⁶ → avoid O(n²)

• n ≥ 10⁷ → need O(n log n) or better

• Use collections.deque, set, dict instead of lists for frequent lookups/removals

12.2 Profiling (cProfile, timeit)

timeit – Quick & accurate timing for small snippets

Python

import timeit # Compare list vs set lookup setup = "data = list(range(1000000))" stmt_list = "999999 in data" stmt_set = "999999 in set(data)" print(timeit.timeit(stmt_list, setup, number=100)) # slow print(timeit.timeit(stmt_set, setup, number=100)) # very fast

cProfile – Full program profiling (find bottlenecks)

Python

import cProfile def slow_function(): total = 0 for i in range(1000000): total += i ** 2 return total cProfile.run("slow_function()")

Output snippet (example):

text

ncalls tottime percall cumtime percall filename:lineno(function) 1000000 0.450 0.000 0.450 0.000 <string>:1(<genexpr>) 1 0.451 0.451 0.451 0.451 <string>:1(slow_function)

Better: Use snakeviz for visualization

Bash

pip install snakeviz python -m cProfile -o profile.out your_script.py snakeviz profile.out

line_profiler – Line-by-line timing (very useful)

Bash

pip install line_profiler

Python

@profile def slow_loop(): total = 0 for i in range(100000): total += i * i return total

Run with:

Bash

kernprof -l script.py python -m line_profiler script.py.lprof

12.3 Efficient Data Structures

Choosing the right structure can give 10×–1000× speedup.

Task / OperationRecommended StructureTime ComplexityWhy? / Alternative (avoid)Frequent lookups / membershipset or dictO(1) avgAvoid list (O(n))Ordered unique itemscollections.OrderedDict or dict (3.7+)O(1)—Fast append/pop from both endscollections.dequeO(1)Avoid list (O(n) for pop(0))Count occurrencescollections.CounterO(n)Avoid manual dict countingPriority queue / min-heapheapqO(log n) push/pop—Sorted list with fast insertionbisect + listO(log n) search, O(n) insertUse when n is smallLarge numerical data / matrix opsnumpy arrayVery fast (C)Avoid Python lists

Example speedup – membership check

Python

import time data_list = list(range(1_000_000)) data_set = set(data_list) start = time.time() 999999 in data_list # O(n) → slow print(time.time() - start) # ~0.1–0.5 sec start = time.time() 999999 in data_set # O(1) → instant print(time.time() - start) # ~0.0000001 sec

12.4 Caching & Memoization

Memoization — cache function results to avoid recomputation.

Built-in: @functools.lru_cache

Python

from functools import lru_cache @lru_cache(maxsize=128) # maxsize=None → unlimited def fibonacci(n): if n < 2: return n return fibonacci(n-1) + fibonacci(n-2) print(fibonacci(35)) # Instant (cached)

Manual cache (simple dict)

Python

def expensive_calc(n, cache={}): if n in cache: return cache[n] result = n 3 + n 2 + n # simulate heavy work cache[n] = result return result

Advanced: @functools.cache (Python 3.9+) Unlimited cache (no maxsize limit)

12.5 NumPy & Pandas for Speed

For numerical/data work — NumPy & Pandas are 10–100× faster than pure Python lists/dicts.

NumPy example – Vectorized operations

Python

import numpy as np # Slow Python loop lst = list(range(1_000_000)) result = [x**2 for x in lst] # ~100 ms # NumPy – blazing fast arr = np.arange(1_000_000) result = arr ** 2 # ~1–5 ms

Pandas for data frames

Python

import pandas as pd # Slow: loop over rows df = pd.DataFrame({"A": range(1000000)}) df["B"] = df["A"] 2 # slow if done with apply() # Fast: vectorized df["B"] = df["A"] 2 # very fast

When to switch to NumPy/Pandas:

• Working with numbers, arrays, matrices → NumPy

• Tabular data, filtering, grouping, CSV/Excel → Pandas

• Avoid loops → use vectorized operations, broadcasting

Mini Project – Speed Comparison Tool

Python

import timeit import numpy as np def python_sum(n): return sum(range(n)) def numpy_sum(n): return np.arange(n).sum() n = 10_000_000 print("Python:", timeit.timeit(lambda: python_sum(n), number=1)) print("NumPy :", timeit.timeit(lambda: numpy_sum(n), number=1))

Output example:

text

Python: 0.45 seconds NumPy : 0.008 seconds

This completes the full Performance Optimization section — now you have the tools to measure bottlenecks, choose efficient structures, and write blazing-fast Python code!

📚 Amazon Book Library

All my books are FREE on Amazon Kindle Unlimited🌍 Exclusive Country-Wise Amazon Book Library – Only Here!

On GlobalCodeMaster.com you’ll find complete, ready-to-use lists of my books with direct Amazon links for every country.

Belong to India, Australia, USA, UK, Canada or any other country? Just click your country’s link and enjoy:

✅ Any eBook FREE on Kindle Unlimited ✅ Or buy at incredibly low prices

400+ fresh books written in 2025-2026 with today’s latest AI, Python, Machine Learning & tech trends – nowhere else will you find this complete country-wise collection on one platform!

Choose your country below and start reading instantly 🚀

BOOK LIBRARY USA 2026 LINK

BOOK LIBRARY INDIA 2026 LINK

BOOK LIBRARY AUSTRALIA 2026 LINK

BOOK LIBRARY CANADA 2026 LINK

BOOK LIBRARY UNITED KINGDOM 2026 LINK

BOOK LIBRARY GERMANY 2026 LINK

BOOK LIBRARY FRANCE 2026 LINK

BOOK LIBRARY ITALY 2026 LINK

BOOK LIBRARY SPAIN 2026 LINK

BOOK LIBRARY NETHERLANDS 2026 LINK

BOOK LIBRARY BRAZIL 2026 LINK

BOOK LIBRARY MEXICO 2026 LINK

BOOK LIBRARY JAPAN 2026 LINK

BOOK LIBRARY POLAND 2026 LINK

BOOK LIBRARY IRELAND 2026 LINK

BOOK LIBRARY SWEDEN 2026 LINK

BOOK LIBRARY BELGIUM 2026 LINK