Javatpoint Logo
Javatpoint Logo

Design a data structure that supports insert, delete, search and getRandom in constant time.

Runtime efficiency is critical for scalability when designing data structures to support essential operations like inserting, deleting, searching, and retrieving elements. As data structures grow to hold more and more details, keeping these operations fast becomes challenging. While primary data structures like arrays or linked lists allow some efficient operations, they fail when it comes to others, especially as they grow huge. An optimum solution is a hybrid data structure that combines the strengths of simpler constructs to enable swift constant-time performance for not just one but all the four critical operations required. In this article, we will discuss a hybrid data structure that pairs a hash table with a dynamic array that can support inserts, deletes, searches and random fetches on massive datasets in guaranteed O(1) time. Developers can craft data containers tailor-made for speed and scale by understanding the synergies between these interlinked formats.

The key highlights are - the need for efficient data structures that can handle big data, the limitations of more superficial data structures, proposing a hash table + array combination for achieving efficiencies in all required operations, and how this hybrid format leverages the strengths of the two constructs.

How to Design the Data Structure?

Hash Table

The hash table can be implemented in Python using a dictionary. Dictionaries provide efficient key-value lookups, inserts and deletes, making them ideal for realizing hash tables.

We can define a dictionary that maps each key to an index value. This index will then link to the element's position in the array. Dictionary operations like setting/getting elements require just O(1) time.

The dictionary size would also ideally be proportional to N (load factor considerations aside) to store N elements. Resizing the underlying dictionary container when it starts to fill up allows the runtime promise to be kept.

The exact mapping of keys to array indices depends on the hash function used. The hash function should provide a uniform distribution of hashes to minimize collisions. Popular choices would include MD5, SHA256 or any pre-existing hashing library.


Python lists serve very well as dynamic arrays. Appending to, inserting in, deleting from and accessing list elements takes time. This meets all our requirements.

The list will store every element inserted into our data structure. By linking indices from the hash table to positions in this list, elements can be efficiently accessed or manipulated at will.

We can start with a list initialized to an appropriate capacity and allow it to expand as needed when it starts to fill up. Choosing the right initial size and expansion parameters can impact performance.

Knowing Operations:

Insert Operation

The insert operation allows the addition of new elements to the data structure. Specifically, it enables the following functionality:

  1. Allocating memory to hold the new element value if capacity allows.
  2. Updating internal data structure tracking to reflect the presence of new elements. This usually involves updating index tables, pointers, etc.
  3. Maintaining insertion order among elements may also be necessary in some cases.

The time complexity for insertion depends on the data structure type and its implementation. The goal is to achieve O(1) constant time inserts irrespective of data structure size.

Delete Operation

The delete operation facilitates the removal of existing elements from the data structure. This entails a few essential tasks:

  1. Locating the element to remove via search lookup, usually based on some identifier or key.
  2. Freeing any memory allocated to the element getting removed.
  3. Updating internal data structure tracking by modifying indices, pointers, etc., to reflect the logical removal of this element.
  4. Any other elements in the data structure may also need to be shifted or relocated to fill the space vacated by element removal and maintain structure continuity.

As with insertion, deletes aim for O(1) time complexity, again independent of the total elements present.

Search Operation

The search functionality allows checking whether a given element currently exists within the data structure. The following are the broad steps involved:

  1. Accept search query with element key or identifier value to lookup.
  2. Scan through data structure tracking metadata like index tables, node pointers, etc., to locate elements.
  3. If a matching element is found, return true or false.

Achieving O(1) lookup time is desirable, though it is still considered efficient up to O(log N) via binary search trees.

getRandom Operation

The getRandom operation fetches an arbitrary element from the data structure uniformly at random. The main steps are:

  1. Generate uniformly random integer index within data structure bounds.
  2. Use the index to access elements stored at that location in memory.
  3. Return accessed random element.

Getting a random element also targets O(1) time, similar to the above operations.

Python Implementation for this DS

This program implements a randomized data structure that supports efficient search, insertion, deletion and random access operations. The structure combines a hash table and dynamic array to store elements randomly, allowing fast lookups and accesses.

Some key capabilities this data structure provides:

  • Fast O(1) search for elements
  • Efficient O(1) removal and insertion of elements
  • Accessing random elements quickly in O(1) time
  • Stores elements in a random order, allowing randomness in access

The random ordering and quick accesses make it suitable for applications needing randomness, like shuffling playlists, games, sampling, and more.

Algorithm Steps

  1. Initialize a hash table (python dictionary) to store key-value pairs mapping elements to their indices
  2. Initialize a dynamic array (python list) to store elements randomly
  3. To insert an element:
    • Check if the element already exists in the hash table
    • If not, append it to the dynamic array
    • Store mapping of the element to its array index in the hash table
  4. To remove an element:
    • Get the index of the element from the hash table
    • Replace the element with the last element in the array
    • Update index mapping for the last element
    • Pop the last element (original copy) from an array
  5. To search for an element:
    • Directly retrieve the index from the hash table
    • Return index if element found else return -1
  6. To get a random element:
    • Use the choice() method to pick a random index
    • Access element at this index in the array


Design a data structure that supports insert, delete, search and getRandom in constant time


The RandomizedDS class implements a data structure that stores elements randomly and allows efficient search, insertion, removal, and random access. Here is a step-by-step explanation:

  1. The init method initializes two data members:
    • - A hash table (python dictionary) that maps each element to its index in the list
    • self.elements - A list containing all the inserted elements
  2. The insert() method:
    • Check if the element already exists using the hash table (O(1) check)
    • If it does not exist, append the element to the end of the list.
    • Updates the hash table by mapping the inserted element to its new index in the list
    • Returns True if the insert was successful, False otherwise
  3. The remove() method:
    • Uses the hash table to retrieve the index of the element to remove in O(1) time
    • Gets the last element from the list and places it at the position of the element to remove
    • Updates the index of the last element in the hash table
    • Pops (removes) the duplicate last element from the end of the list
    • Deletes the key to be removed from the hash table
  4. The getRandom() method picks an element randomly from the list using random.choice()
  5. The search() method uses the hash table to retrieve the element's index in O(1) time. If the element exists, else returns -1
  6. Example usage:
    1. Create a RandomizedDS object
    2. Insert elements
    3. Search for inserted elements (True indicates found)
    4. Print randomly picked element
    5. Try removing the element
    6. Search again to check removal


In simpler terms, we saw how mixing two primary data structures - hash tables and arrays - can give us a customized container that performs excellent across the board. By exploiting what each one does best, we get a fast all-rounder.

Hash tables use clever numbering to access data entries directly. Arrays place items sequentially to allow easy inserts and random picks. Combining them covers holes in each one's capabilities. The joint structure gives us speedy addition, removal, finding, and randomly getting elements, even in huge collections.

The techniques we need are also simple to grasp at a high level. Hash functions map keys to array spots in a balanced manner. Reserving extra space avoids crowding, which slows things down. While coding natural systems using these ideas adds complexity, the concepts are intuitive.

We specifically looked at guidelines for building a customizable data store in Python. Its standard dictionary and list types already supply ingredients required for high efficiency. Simply glueing them together correctly allows versatile structures to be crafted with little effort.

In the data analytics domain, such customizable containers form building bricks. Having strong guarantees of speed despite vast data sizes unlocks scalable architectures. Innovative products benefit end users by providing responsive storage, retrieval and sharing of information.

Understanding these basic techniques to create data structures tailored to needs is critical for engineers working on analytics pipelines.

The article showed how combining complementary approaches gives customizable and efficient solutions that are more significant than the sum of parts. Analytics systems serving real-world demands can be built using these Lego blocks.

Youtube For Videos Join Our Youtube Channel: Join Now


Help Others, Please Share

facebook twitter pinterest

Learn Latest Tutorials


Trending Technologies

B.Tech / MCA