Guide to Hash Tables in Python

Trending 5 months ago


Hash tables connection an businesslike and elastic method of storing and retrieving data, making them indispensable for tasks involving ample information sets aliases requiring accelerated entree to stored items.

While Python doesn't person a built-in information building explicitly called a "hash table", it provides nan dictionary, which is simply a shape of a hash table. Python dictionaries are unordered collections of key-value pairs, wherever nan cardinal is unsocial and holds a corresponding value. Thanks to a process known arsenic "hashing", dictionaries alteration businesslike retrieval, addition, and removal of entries.

Note: If you're a Python programmer and person ever utilized a dictionary to shop information arsenic key-value pairs, you've already benefited from hash array exertion without needfully knowing it! Python dictionaries are implemented utilizing hash tables!

In this guide, we'll delve into nan world of hash tables. We'll commencement pinch nan basics, explaining what hash tables are and really they work. We'll besides research Python's implementation of hash tables via dictionaries, supply a step-by-step guideline to creating a hash array successful Python, and moreover touch connected really to grip hash collisions. Along nan way, we'll show nan inferior and ratio of hash tables pinch real-world examples and useful Python snippets.

Defining Hash Tables: Key-Value Pair Data Structure

Since dictionaries successful Python are fundamentally an implementation of hash tables, let's first attraction connected what hash tables really are, and dive into Python implementation afterward.

Hash tables are a type of information building that provides a system to shop information successful an associative manner. In a hash table, information is stored successful an array format, but each information worth has its ain unique key, which is utilized to place nan data. This system is based connected key-value pairs, making nan retrieval of information a swift process.

The affinity often utilized to explicate this conception is simply a real-world dictionary. In a dictionary, you usage a known connection (the "key") to find its meaning (the "value"). If you cognize nan word, you tin quickly find its definition. Similarly, successful a hash table, if you cognize nan key, you tin quickly retrieve its value.

Essentially, we are trying to shop key-value pairs successful nan astir businesslike measurement possible.

For example, opportunity we want to create a hash array that stores nan commencement period of various people. The people's names are our keys and their commencement months are nan values:

+-----------------------+ | Key | Value | +-----------------------+ | Alice | January | | Bob | May | | Charlie | January | | David | August | | Eve | December | | Brian | May | +-----------------------+

To shop these key-value pairs successful a hash table, we'll first request a measurement to person nan worth of keys to nan due indexes of nan array that represents a hash table. That's wherever a hash function comes into play! Being nan backbone of a hash array implementation, this usability processes nan cardinal and returns nan corresponding scale successful nan information retention array - conscionable arsenic we need.

The extremity of a good hash function is to administer nan keys evenly crossed nan array, minimizing nan chance of collisions (where 2 keys nutrient nan aforesaid index).

In reality, hash functions are overmuch much complex, but for simplicity, let's usage a hash usability that maps each sanction to an scale by taking nan ASCII worth of nan first missive of nan sanction modulo nan size of nan table:

def simple_hash(key, array_size): return ord(key[0]) % array_size

This hash usability is simple, but it could lead to collisions because different keys mightiness commencement pinch nan aforesaid missive and hence nan resulting indices will beryllium nan same. For example, opportunity our array has nan size of 10, moving nan simple_hash(key, 10) for each of our keys will springiness us:

Alternatively, we tin reshape this information successful a much concise way:

+---------------------+ | Key | Index | +---------------------+ | Alice | 5 | | Bob | 6 | | Charlie | 7 | | David | 8 | | Eve | 9 | | Brian | 6 | +---------------------+

Here, Bob and Brian person nan aforesaid scale successful nan resulting array, which results successful a collision. We'll talk much astir collisions successful nan second sections - some successful position of creating hash functions that minimize nan chance of collisions and resolving collisions erstwhile they occur.

Designing robust hash functions is 1 of nan astir important aspects of hash array efficiency!

Note: In Python, dictionaries are an implementation of a hash table, wherever nan keys are hashed, and nan resulting hash worth determines wherever successful nan dictionary's underlying information retention nan corresponding worth is placed.

In nan pursuing sections, we'll dive deeper into nan soul workings of hash tables, discussing their operations, imaginable issues (like collisions), and solutions to these problems.

Demystifying nan Role of Hash Functions successful Hash Tables

Hash functions are nan heart and soul of hash tables. They service arsenic a span betwixt nan keys and their associated values, providing a intends of efficiently storing and retrieving data. Understanding nan domiciled of hash functions successful hash tables is important to grasp really this powerful information building operates.

What is simply a Hash Function?

In nan discourse of hash tables, a hash usability is simply a typical usability that takes a key arsenic input and returns an index which nan corresponding worth should beryllium stored aliases retrieved from. It transforms nan cardinal into a hash - a number that corresponds to an scale successful nan array that forms nan underlying building of nan hash table.

The hash usability needs to beryllium deterministic, meaning that it should ever nutrient nan aforesaid hash for nan aforesaid key. This way, whenever you want to retrieve a value, you tin usage nan hash usability connected nan cardinal to find retired wherever nan worth is stored.

The Role of Hash Functions successful Hash Tables

The main nonsubjective of a hash usability successful a hash array is to administer nan keys as uniformly arsenic possible crossed nan array. This is important because nan azygous distribution of keys allows for a changeless clip complexity of O(1) for information operations specified arsenic insertions, deletions, and retrievals on average.

To exemplify really a hash usability useful successful a hash table, let's again return a look astatine nan illustration from nan erstwhile section:

+-----------------------+ | Key | Value | +-----------------------+ | Alice | January | | Bob | May | | Charlie | January | | David | August | | Eve | December | | Brian | May | +-----------------------+

As before, presume we person a hash function, simple_hash(key), and a hash array of size 10.

As we've seen before, running, say, "Alice" done nan simple_hash() usability returns nan scale 5. That intends we tin find nan constituent pinch nan cardinal "Alice" and nan worth "January" successful nan array representing nan hash table, connected nan scale 5 (6th constituent of that array):

And that applies to each cardinal of our original data. Running each cardinal done nan hash usability will springiness america nan integer worth - an scale successful nan hash array array wherever that constituent is stored:

+---------------------+ | Key | Index | +---------------------+ | Alice | 5 | | Bob | 6 | | Charlie | 7 | | David | 8 | | Eve | 9 | | Brian | 6 | +---------------------+

This tin easy construe to nan array representing a hash array - an constituent pinch nan cardinal "Alice" will beryllium stored nether scale 5, "Bob" nether scale 6, and truthful on:

Note: Under nan scale 6 location are 2 elements - {"Bob", "February"} and {"Brian", "May"}. In nan illustration above, that collision was solved utilizing nan method called separate chaining, which we'll talk astir much later successful this article.

When we want to retrieve nan worth associated pinch nan cardinal "Alice", we again walk nan cardinal to nan hash function, which returns nan scale 5. We past instantly entree nan worth astatine scale 3 of nan hash table, which is "January".

Challenges pinch Hash Functions

While nan thought down hash functions is reasonably straightforward, designing a bully hash usability tin beryllium challenging. A superior interest is what's known arsenic a collision, which occurs erstwhile 2 different keys hash to nan aforesaid scale successful nan array.

Just return a look astatine nan "Bob" and "Brian" keys successful our example. They person nan aforesaid index, meaning they are stored successful nan aforesaid spot successful nan hash array array. In its essence, this is an illustration of a hashing collision.

The likelihood of collisions is dictated by nan hash usability and nan size of nan hash table. While it's virtually intolerable to wholly debar collisions for immoderate non-trivial magnitude of data, a bully hash usability coupled pinch an appropriately sized hash array will minimize nan chances of collisions.

Different strategies specified arsenic open addressing and separate chaining tin beryllium utilized to resoluteness collisions erstwhile they occur, which we'll screen successful a later section.

Analyzing Time Complexity of Hash Tables: A Comparison

One of nan cardinal benefits of utilizing hash tables, which sets them isolated from galore different information structures, is their clip complexity for basal operations. Time complexity is simply a computational conception that refers to nan magnitude of clip an cognition aliases a usability takes to run, arsenic a usability of nan size of nan input to nan program.

When discussing clip complexity, we mostly mention to 3 cases:

  1. Best Case: The minimum clip required for executing an operation.
  2. Average Case: The mean clip needed for executing an operation.
  3. Worst Case: The maximum clip needed for executing an operation.

Hash tables are particularly noteworthy for their impressive clip complexity successful nan average case scenario. In that scenario, basal operations successful hash tables (inserting, deleting, and accessing elements) person a constant clip complexity of O(1).

The changeless clip complexity implies that nan clip taken to execute these operations remains constant, sloppy of nan number of elements successful nan hash table.

This makes these operations highly efficient, particularly erstwhile dealing pinch ample datasets.

While nan mean lawsuit clip complexity for hash tables is O(1), the worst-case script is simply a different story. If aggregate keys hash to nan aforesaid scale (a business known arsenic a collision), nan clip complexity tin degrade to O(n), wherever n is nan number of keys mapped to nan aforesaid index.

This script occurs because, erstwhile resolving collisions, further steps must beryllium taken to shop and retrieve data, typically by traversing a linked database of entries that hash to nan aforesaid index.

Note: With a well-designed hash usability and a correctly sized hash table, this worst-case script is mostly nan objection alternatively than nan norm. A bully hash usability paired pinch due collision solution strategies tin support collisions to a minimum.

Comparing to Other Data Structures

When compared to different information structures, hash tables guidelines retired for their efficiency. For instance, operations for illustration search, insertion, and deletion successful a balanced binary hunt character aliases a balanced AVL Tree person a clip complexity of O(log n), which, though not bad, is not arsenic businesslike arsenic nan O(1) clip complexity that hash tables connection successful nan mean case.

While arrays and linked lists connection O(1) clip complexity for immoderate operations, they can't support this level of ratio crossed each basal operations. For example, searching successful an unsorted array aliases linked database takes O(n) time, and insertion successful an array takes O(n) clip successful nan worst case.

Python's Approach to Hash Tables: An Introduction to Dictionaries

Python provides a built-in information building that implements nan functionality of a hash array called a dictionary, often referred to arsenic a "dict". Dictionaries are 1 of Python's astir powerful information structures, and knowing really they activity tin importantly boost your programming skills.

Check retired our hands-on, applicable guideline to learning Git, pinch best-practices, industry-accepted standards, and included cheat sheet. Stop Googling Git commands and really learn it!

What are Dictionaries?

In Python, dictionaries (dicts) are unordered collections of key-value pairs. Keys successful a dictionary are unsocial and immutable, which intends they can't beryllium changed erstwhile they're set. This spot is basal for nan correct functioning of a hash table. Values, connected nan different hand, tin beryllium of immoderate type and are mutable, meaning you tin alteration them.

A key-value brace successful a dictionary is besides known arsenic an item. Each cardinal successful a dictionary is associated (or mapped) to a azygous value, forming a key-value pair:

my_dict = {"Alice": "January", "Bob": "May", "Charlie": "January"}

How do Dictionaries Work successful Python?

Behind nan scenes, Python's dictionaries run arsenic a hash table. When you create a dictionary and adhd a key-value pair, Python applies a hash usability to nan key, which results successful a hash value. This hash worth past determines wherever successful representation nan corresponding worth will beryllium stored.

The beauty of this is that erstwhile you want to retrieve nan value, Python applies nan aforesaid hash usability to nan key, which quickly guides Python to wherever nan worth is stored, sloppy of nan size of nan dictionary:

my_dict = {} my_dict["Alice"] = "January" print(my_dict["Alice"])

Key Operations and Time Complexity

Python's built-in dictionary information building makes performing basal hash array operations—such arsenic insertions, access, and deletions a breeze. These operations typically person an mean clip complexity of O(1), making them remarkably efficient.

Note: As pinch hash tables, nan worst-case clip complexity tin beryllium O(n), but this happens rarely, only erstwhile location are hash collisions.

Inserting key-value pairs into a Python dictionary is straightforward. You simply delegate a worth to a cardinal utilizing nan duty usability (=). If nan cardinal doesn't already beryllium successful nan dictionary, it's added. If it does exist, its existent worth is replaced pinch nan caller value:

my_dict = {} my_dict["Alice"] = "January" my_dict["Bob"] = "May" print(my_dict)

Accessing a value successful a Python dictionary is conscionable arsenic elemental arsenic inserting one. You tin entree nan worth associated pinch a peculiar cardinal by referencing nan cardinal successful quadrate brackets. If you effort to entree a cardinal that doesn't beryllium successful nan dictionary, Python will raise a KeyError:

print(my_dict["Alice"]) print(my_dict["Charlie"])

To forestall this error, you tin usage nan dictionary's get() method, which allows you to return a default worth if nan cardinal doesn't exist:

print(my_dict.get("Charlie", "Unknown"))

Note: Similarly, nan setdefault() method tin beryllium utilized to safely insert a key-value brace into nan dictionary if nan cardinal doesn't already exist:

my_dict.setdefault("new_key", "default_value")

You tin remove a key-value pair from a Python dictionary utilizing nan del keyword. If nan cardinal exists successful nan dictionary, it's removed on pinch its value. If nan cardinal doesn't exist, Python will besides raise a KeyError:

del my_dict["Bob"] print(my_dict) del my_dict["Bob"]

Like pinch access, if you want to forestall an correction erstwhile trying to delete a cardinal that doesn't exist, you tin usage nan dictionary's pop() method, which removes a key, returns its worth if it exists, and returns a default worth if it doesn't:

print(my_dict.pop("Bob", "Unknown"))

All-in-all, Python dictionaries service arsenic a high-level, user-friendly implementation of a hash table. They are easy to use, yet powerful and efficient, making them an fantabulous instrumentality for handling a wide assortment of programming tasks.

Advice: If you're testing for rank (i.e., whether an point is successful a collection), a dictionary (or a set) is often a much businesslike prime than a database aliases a tuple, particularly for larger collections. That's because dictionaries and sets usage hash tables, which let them to trial for rank successful changeless clip (O(1)), arsenic opposed to lists aliases tuples, which return linear clip (O(n)).

In nan adjacent sections, we will dive deeper into nan applicable aspects of utilizing dictionaries successful Python, including creating dictionaries (hash tables), performing operations, and handling collisions.

How to Create Your First Hash Table successful Python

Python's dictionaries supply a ready-made implementation of hash tables, allowing you to shop and retrieve key-value pairs pinch fantabulous efficiency. However, to understand hash tables thoroughly, it tin beryllium beneficial to instrumentality 1 from scratch. In this section, we'll guideline you done creating a elemental hash array successful Python.

We'll commencement by defining a HashTable class. The hash array will beryllium represented by a database (the table), and we will usage a very elemental hash usability that calculates nan remainder of nan ASCII worth of nan cardinal string's first characteristic divided by nan size of nan table:

class HashTable: def __init__(self, size): self.size = size self.table = [None]*size def _hash(self, key): return ord(key[0]) % self.size

In this class, we person nan __init__() method to initialize nan hash table, and a _hash() method, which is our elemental hash function.

Now, we'll adhd methods to our HashTable people for adding key-value pairs, getting values by key, and removing entries:

class HashTable: def __init__(self, size): self.size = size self.table = [None]*size def _hash(self, key): return ord(key[0]) % self.size def set(self, key, value): hash_index = self._hash(key) self.table[hash_index] = (key, value) def get(self, key): hash_index = self._hash(key) if self.table[hash_index] is not None: return self.table[hash_index][1] raise KeyError(f'Key {key} not found') def remove(self, key): hash_index = self._hash(key) if self.table[hash_index] is not None: self.table[hash_index] = None else: raise KeyError(f'Key {key} not found')

The set() method adds a key-value brace to nan table, while nan get() method retrieves a worth by its key. The remove() method deletes a key-value brace from nan hash table.

Note: If nan cardinal doesn't exist, nan get and region methods raise a KeyError.

Now, we tin create a hash array and usage it to shop and retrieve data:

hash_table = HashTable(10) hash_table.set('Alice', 'January') hash_table.set('Bob', 'May') print(hash_table.get('Alice')) hash_table.remove('Bob') print(hash_table.get('Bob'))

Note: The supra hash array implementation is rather elemental and does not grip hash collisions. In real-world use, you'd request a much blase hash usability and collision solution strategy.

Resolving Collisions successful Python Hash Tables

Hash collisions are an inevitable portion of utilizing hash tables. A hash collision occurs erstwhile 2 different keys hash to nan aforesaid scale successful nan hash table. As Python dictionaries are an implementation of hash tables, they besides request a measurement to grip these collisions.

Python's built-in hash array implementation uses a method called "open addressing" to grip hash collisions. However, to amended understand nan collision solution process, let's talk a simpler method called "separate chaining".

Separate Chaining

Separate chaining is simply a collision solution method successful which each slot successful nan hash array holds a linked database of key-value pairs. When a collision occurs (i.e., 2 keys hash to nan aforesaid index), nan key-value brace is simply appended to nan extremity of nan linked database astatine nan colliding index.

Remember, we had a collision successful our illustration because some "Bob" and "Brian" had nan aforesaid scale - 6. Let's usage that illustration to exemplify nan system down abstracted chaining. If we were to presume that nan "Bob" constituent was added to nan hash array first, we'd tally into nan problem erstwhile trying to shop nan "Brian" constituent since nan scale 6 was already taken.

Solving this business utilizing abstracted chaining would see adding nan "Brian" constituent arsenic nan 2nd constituent of nan linked database assigned to scale 6 (the "Bob" constituent is nan first constituent of that list). And that's each location is to it, conscionable arsenic shown successful nan pursuing illustration:

Here's really we mightiness modify our HashTable people from nan erstwhile illustration to usage abstracted chaining:

class HashTable: def __init__(self, size): self.size = size self.table = [[] for _ in range(size)] def _hash(self, key): return ord(key[0]) % self.size def set(self, key, value): hash_index = self._hash(key) for kvp in self.table[hash_index]: if kvp[0] == key: kvp[1] = value return self.table[hash_index].append([key, value]) def get(self, key): hash_index = self._hash(key) for kvp in self.table[hash_index]: if kvp[0] == key: return kvp[1] raise KeyError(f'Key {key} not found') def remove(self, key): hash_index = self._hash(key) for i, kvp in enumerate(self.table[hash_index]): if kvp[0] == key: self.table[hash_index].pop(i) return raise KeyError(f'Key {key} not found')

In this updated implementation, nan array is initialized arsenic a database of quiet lists (i.e., each slot is an quiet linked list). In nan set() method, we iterate complete nan linked database astatine nan hashed index, updating nan worth if nan cardinal already exists. If it doesn't, we append a caller key-value brace to nan list.

The get() and remove() methods besides request to iterate complete nan linked database astatine nan hashed scale to find nan cardinal they're looking for.

While this attack solves nan problem of collisions, it does lead to an summation successful clip complexity erstwhile collisions are frequent.

Open Addressing

The method utilized by Python dictionaries to grip collisions is much blase than abstracted chaining. Python uses a shape of unfastened addressing called "probing".

In probing, erstwhile a collision occurs, nan hash array checks nan adjacent disposable slot and places nan key-value brace location instead. The process of uncovering nan adjacent disposable slot is called "probing", and respective strategies tin beryllium used, specified as:

  • Linear probing - checking 1 slot astatine a clip successful order
  • Quadratic probing - checking slots successful expanding powers of two

Note: The circumstantial method Python uses is much analyzable than immoderate of these, but it ensures that lookups, insertions, and deletions stay adjacent to O(1) clip complexity moreover successful cases wherever collisions are frequent.

Let's conscionable return a speedy look astatine nan collision illustration from nan erstwhile section, and show really would we dainty it utilizing nan unfastened addressing method. Say we person a hash array pinch only 1 constituent - {"Bob", "May"} connected nan scale number 6. Now, we wouldn't beryllium capable to adhd nan "Brian" constituent to nan hash array owed to nan collision. But, nan system of linear probing tells america to shop it successful nan first quiet scale - 7. That's it, easy right?


From their conceptual underpinnings to their implementation arsenic dictionaries successful Python, hash tables guidelines arsenic 1 of nan astir powerful and versatile information structures. They let america to efficiently store, retrieve, and manipulate information successful our programs, making them invaluable for a multitude of real-world applications specified arsenic caching, information indexing, wave analysis, and overmuch more.

Hash tables beryllium their prowess to their clip complexity of O(1) for basal operations, making them exceptionally accelerated moreover pinch ample amounts of data. Moreover, their collision solution strategies, specified arsenic Python's unfastened addressing approach, guarantee that they negociate collisions effectively, maintaining their efficiency.

While dictionaries, arsenic Python's implementation of hash tables, are powerful and efficient, they do devour more representation than different information structures for illustration lists aliases tuples. This is mostly a adjacent trade-off for nan capacity benefits they offer, but if representation usage is simply a interest (for instance, if you're moving pinch a very ample dataset), it's thing to support successful mind.

In specified cases, you whitethorn want to see alternatives for illustration lists of tuples for mini datasets aliases much memory-efficient information structures provided by libraries for illustration NumPy aliases pandas for larger datasets.

Source Stack Abuse
Stack Abuse