5 less known Python features. Each data scientist should know

Photo via editor Chatgpt

# Entry

Python It is one of the most popular languages ​​used in the sphere of Data Science, valued for its simplicity, versatility and a powerful library ecosystem, including NumbersIN PandyIN Scikit-LearnAND Tensorflow. While these tools provide a significant part of bulky lift, Python itself contains a number of functions that can support you write a cleaner, faster and more capable code. Many of these possibilities become unnoticed, but they can improve the way in which the structure and manage your projects.

In this article we examine five less known but favorable features of Python, which every scientist should have tools in their set.

# 1 else Loop clause

Do you know for AND while Python loops may have else clause?

Although this may seem contrary to intuition at the beginning, else The block is made only when the loop ends without break statement. This is useful when searching the data set and you want to start some logic only if a specific condition has never been met.

for row in dataset:
    if row['target'] == 'desired_value':
        print("Found!")
        break
else:
    print("Not found.")

In this fragment, else The block is made only when the loop ends without encountering a break. This avoids creating additional flags or conditions outside the loop.

# 2 dataclasses Module

. Data classes The module, introduced in Python 3.7, provides the functions of a decorator and a midfielder that automatically generate special methods such as __init__()IN __repr__()AND __eq__() for your classes. This is useful in learning data when you need featherlight classes for storing parameters, results or configuration settings without writing a repetitive code from the boiler.

from dataclasses import dataclass

@dataclass
class ExperimentConfig:
    learning_rate: float
    batch_size: int
    epochs: int

WITH @dataclassYou get a tidy designer, a clear chain representation and comparative possibilities.

# 3. Walrus operator (:=)

. Walrus operator (:=), introduced in Python 3.8, allows you to assign values ​​to variables as part of the expression. This is useful when you want to calculate and test the value without repeating calculations in many places.

data = [1, 2, 3, 4, 5]

if (avg := sum(data) / len(data)) > 3:
    print(f"Average is {avg}")

Here, avg It is assigned and checked at the same time. This removes the need for a different line and makes it easier to read the code.

# 4. enumerate() for indexed loops

When you need both the index and values ​​during iteration, enumerate() This is the most pitty way to do this. Requires any iterative (like a list, brief or string) and returns steam (index, value) during the loop.

for i, row in enumerate(data):
    print(f"Row {i}: {row}")

This improves readability, reduces the chances of errors and makes your intention clearer. This is useful in learning data when items of data of data or results with items that matter.

# 5. collections Module

Python collections The module provides specialized container data that can be more capable and expressive than using only lists or dictionaries. Among the most popular is Counterwhich can count elements in iteral with a minimum code.

from collections import Counter

word_counts = Counter(words)
most_common = word_counts.most_common(5)

Do you need an arranged dictionary? Apply OrderedDict. Do you need a default dictionary? To try defaultdict. These tools eliminate the need for manual manual logic and can even improve the performance in enormous -scale data processing.

# Application

Tools such as else loop clause, dataclassesAnd the Walrus operator can eliminate the unnecessary boiler plate and make logic more concise. Functions such as enumerate() and modules such as collections Support ite, count and organize data with elegance and performance. Taking into account these less known jewels for work flow, you can reduce the complexity, avoid typical traps and focus more on solving a real problem with data, not on creating code.

Jayita Gulati She is an enthusiast of machine learning and a technical writer driven by her passion for building machine learning models. He has a master’s degree in computer science at the University of Liverpool.