Steven Stuart logo

Search Algorithms

📖 4 min read

Why Searching Exists

Finding specific data efficiently is fundamental to most applications. The choice of search algorithm can mean the difference between instant response and waiting minutes for results.

Use when:

  • Data is unsorted and can’t be sorted
  • Very small datasets (<20 items)
  • Searching for multiple matches
  • Simple one-time searches
  • Need to find all occurrences of a value

Don’t use when:

  • Data can be sorted (use binary search instead)
  • Searching repeatedly on same dataset
  • Large datasets (>1000 items)
  • Performance is critical

Modern reality: Built into language methods like C#’s Contains(), IndexOf(), and LINQ methods. Understand the concept but use built-ins in production.

Time Complexity

  • Best case: O(1) - element is first
  • Average case: O(n/2) = O(n)
  • Worst case: O(n) - element is last or not found
  • Space complexity: O(1)

C# Implementation

public static class LinearSearch
{
    public static int Search<T>(T[] searchList, T targetValue) where T : IEquatable<T>
    {
        for (int idx = 0; idx < searchList.Length; idx++)
        {
            Console.WriteLine($"Checking index {idx}: {searchList[idx]}");
            if (searchList[idx].Equals(targetValue))
            {
                Console.WriteLine($"Found {targetValue} at index {idx}");
                return idx;
            }
        }
        
        throw new ArgumentException($"{targetValue} not found in list");
    }
    
    public static List<int> SearchAll<T>(T[] searchList, T targetValue) where T : IEquatable<T>
    {
        var indices = new List<int>();
        for (int idx = 0; idx < searchList.Length; idx++)
        {
            if (searchList[idx].Equals(targetValue))
            {
                indices.Add(idx);
            }
        }
        return indices;
    }
}

// Example usage
string[] recipe = {"nori", "tuna", "soy sauce", "sushi rice", "tuna"};

try
{
    int index = LinearSearch.Search(recipe, "tuna");
    Console.WriteLine($"First occurrence of 'tuna' at index {index}");
}
catch (ArgumentException e)
{
    Console.WriteLine(e.Message);
}

// Find all occurrences
var allIndices = LinearSearch.SearchAll(recipe, "tuna");
Console.WriteLine($"All occurrences of 'tuna': [{string.Join(", ", allIndices)}]");

Use when:

  • Data is sorted (or can be sorted once)
  • Searching repeatedly on same dataset
  • Large datasets where O(log n) vs O(n) matters
  • Need guaranteed logarithmic performance

Don’t use when:

  • Data is unsorted and can’t be sorted
  • One-time searches on small data
  • Frequent insertions/deletions (ruins sorted property)

Modern reality: Implement this yourself in interviews. Use built-ins like Array.BinarySearch() and List<T>.BinarySearch() in C# for production.

Time Complexity

  • All cases: O(log n)
  • Space complexity: O(1) iterative, O(log n) recursive (call stack)

Binary search works by repeatedly dividing the search space in half. Each comparison eliminates half of the remaining elements.

Key Insight

Binary search works by repeatedly dividing the search space in half. Each comparison eliminates half of the remaining elements.

C# Implementation

public static class BinarySearch
{
    // Recursive version
    public static int SearchRecursive<T>(T[] sortedList, int leftPointer, int rightPointer, T target) 
        where T : IComparable<T>
    {
        // Base case: search space is empty
        if (leftPointer >= rightPointer)
            return -1;
        
        // Find middle point
        int midIdx = (leftPointer + rightPointer) / 2;
        T midVal = sortedList[midIdx];
        
        Console.WriteLine($"Range [{leftPointer}:{rightPointer}], mid={midIdx}, value={midVal}");
        
        int comparison = midVal.CompareTo(target);
        if (comparison == 0)
            return midIdx;
        else if (comparison > 0)
            // Target is in left half
            return SearchRecursive(sortedList, leftPointer, midIdx, target);
        else
            // Target is in right half
            return SearchRecursive(sortedList, midIdx + 1, rightPointer, target);
    }
    
    // Iterative version (preferred)
    public static int SearchIterative<T>(T[] sortedList, T target) where T : IComparable<T>
    {
        int leftPointer = 0;
        int rightPointer = sortedList.Length;
        
        while (leftPointer < rightPointer)
        {
            int midIdx = (leftPointer + rightPointer) / 2;
            T midVal = sortedList[midIdx];
            
            Console.WriteLine($"Range [{leftPointer}:{rightPointer}], mid={midIdx}, value={midVal}");
            
            int comparison = midVal.CompareTo(target);
            if (comparison == 0)
                return midIdx;
            else if (comparison > 0)
                rightPointer = midIdx;
            else
                leftPointer = midIdx + 1;
        }
        
        return -1; // Not found
    }
}

// Example usage
int[] values = {77, 80, 102, 123, 288, 300, 540};
int result = BinarySearch.SearchRecursive(values, 0, values.Length, 288);
Console.WriteLine($"Element 288 found at index: {result}");

// Test cases
int[] testArray = {5, 6, 7, 8, 9};
Console.WriteLine($"Search for 9: {BinarySearch.SearchIterative(testArray, 9)}");   // 4
Console.WriteLine($"Search for 10: {BinarySearch.SearchIterative(testArray, 10)}"); // -1
Console.WriteLine($"Search for 8: {BinarySearch.SearchIterative(testArray, 8)}");   // 3

Binary Search Variations

1. Find First Occurrence

2. Find Last Occurrence

3. Search Insert Position

Common Interview Problems

1. Search in Rotated Sorted Array

2. Find Peak Element

Performance Comparison

Search Performance on Large Dataset

Modern Usage

C#:

  • Use Array.BinarySearch() for arrays
  • Use List<T>.BinarySearch() for lists
  • Use Contains() for membership testing
  • Use LINQ methods like Where(), First(), Any() for complex searches
  • Use IndexOf() and FindIndex() for finding positions

Interview Tip

Always ask if data is sorted before choosing a search algorithm. This single question determines whether you should use O(n) linear search or O(log n) binary search.

Key Interview Points

  1. Always ask if data is sorted: This determines algorithm choice

    Quick Reference

Search Algorithm Comparison

| Algorithm | Time | Space | Prerequisites | Best For | |———–|——|——-|—————|———-| | Linear Search | O(n) | O(1) | None | Small/unsorted data | | Binary Search | O(log n) | O(1) | Sorted array | Large sorted data | | Hash Table | O(1) avg | O(n) | Good hash function | Key-value lookups |

Decision Flow

Is data sorted?
├─ Yes → Binary Search O(log n)
└─ No → Multiple searches?
   ├─ Yes → Hash Table O(1) per search
   └─ No → Linear Search O(n)

Binary Search Variants

  • Standard: Find any occurrence
  • First/Last: Find boundary of duplicates
  • Insert Position: Where to insert to maintain sort
  • Rotated Array: Find in sorted-rotated array

Found this guide helpful? Share it with your team:

Share on LinkedIn