Skip to content
Tutorial Reference

How to Find Second Smallest or Largest Number in a List in Python

Learn Python methods to find the second smallest or largest number in a list, handling duplicates and edge cases with code examples for sorting, heapq, and manual iteration.

Finding the second smallest or second largest value in a list is a common programming challenge that goes beyond simply using min() or max(). This often arises in data analysis, ranking problems, or algorithm exercises.

This guide explores several Python techniques to find the second smallest and second largest numbers, discussing how different methods handle duplicate values and edge cases.

Understanding "Second" Smallest/Largest (Handling Duplicates)

Before implementing, clarify what "second smallest" or "second largest" means, especially when duplicate values exist:

  • Second Unique Smallest/Largest: Ignores duplicates. In [1, 1, 3, 5], the second unique smallest is 3. In [10, 8, 8, 5], the second unique largest is 8.
  • Second Element After Sorting: Considers duplicates. If [1, 1, 3, 5] is sorted, the element at index 1 (the second position) is 1. If [10, 8, 8, 5] is sorted ([5, 8, 8, 10]), the element at index -2 (second from the end) is 8.

Different methods below naturally align with one of these interpretations. Choose the method that matches your specific requirement.

Finding the Second Smallest Number

Method 1: Using set, sorted, and Indexing (Unique Values)

This method finds the second unique smallest value.

def second_smallest_unique_sorted(data_list):
  """Finds the second unique smallest number using set and sorted."""
  if not data_list:
    return None # Handle empty list

  unique_elements = set(data_list)
  if len(unique_elements) < 2:
    return None # Not enough unique elements

  return sorted(unique_elements)[1]

list1 = [3, 1, 4, 1, 5, 9, 2, 6]            # Unique sorted: [1, 2, 3, 4, 5, 6, 9] -> 2nd is 2
list2 = [1, 1, 1, 3, 5]                     # Unique sorted: [1, 3, 5] -> 2nd is 3
list3 = [5, 5]                              # Only one unique element
list4 = []                                  # Empty list

print(second_smallest_unique_sorted(list1)) # Output: 2
print(second_smallest_unique_sorted(list2)) # Output: 3
print(second_smallest_unique_sorted(list3)) # Output: None
print(second_smallest_unique_sorted(list4)) # Output: None
  • set(data_list): Creates a set, automatically removing duplicate values.
  • sorted(...): Sorts the unique elements in ascending order.
  • [1]: Accesses the element at index 1, which is the second unique smallest value.
  • Checks are added for empty lists and lists with fewer than two unique elements.

Method 2: Using heapq.nsmallest()

The heapq module provides efficient ways to find the N smallest/largest elements. This method considers duplicates similarly to sorting.

import heapq # Required import

def second_smallest_heapq(data_list):
  """Finds the second smallest number using heapq.nsmallest()."""
  if not data_list or len(data_list) < 2:
      try:
          if len(data_list) == 1: return None
          smallest_two = heapq.nsmallest(2, data_list)
          return smallest_two[1] if len(smallest_two) >= 2 else None
      except IndexError: # Should be caught by len checks, but for safety
          return None
  else:
    return heapq.nsmallest(2, data_list)[1]


list1 = [3, 1, 4, 1, 5, 9, 2, 6]    # Smallest two are [1, 1] -> 2nd is 1
list2 = [10, 5, 20, 15]             # Smallest two are [5, 10] -> 2nd is 10
list3 = [1, 1, 1, 3]                # Smallest two are [1, 1] -> 2nd is 1
list4 = [5]                         # Only one element

print(second_smallest_heapq(list1)) # Output: 1
print(second_smallest_heapq(list2)) # Output: 10
print(second_smallest_heapq(list3)) # Output: 1
print(second_smallest_heapq(list4)) # Output: None
  • heapq.nsmallest(2, data_list): Efficiently finds the two smallest elements in the list, returning them as a new list. It handles duplicates like sorting does.
  • [1]: Accesses the second element (index 1) of the result list.
  • Edge case handling for lists shorter than 2 elements is added.

Method 3: Manual Iteration (Looping)

This method finds the second unique smallest value by keeping track of the two smallest distinct numbers seen so far.

def second_smallest_loop(data_list):
  """Finds the second unique smallest number using a loop."""
  if not data_list or len(set(data_list)) < 2: # Need at least 2 unique elements
      return None

  smallest = float('inf')
  second_smallest_val = float('inf')

  for number in data_list:
    if number < smallest:
      second_smallest_val = smallest    # Old smallest becomes second smallest
      smallest = number                 # Update smallest
    elif number < second_smallest_val and number != smallest:
      second_smallest_val = number

  return second_smallest_val if second_smallest_val != float('inf') else None

list1 = [3, 1, 4, 1, 5, 9, 2, 6]    # -> 2
list2 = [1, 1, 1, 3, 5]             # -> 3
list3 = [5, 5]                      # -> None
list4 = [10]                        # -> None

print(second_smallest_loop(list1))  # Output: 2
print(second_smallest_loop(list2))  # Output: 3
print(second_smallest_loop(list3))  # Output: None
print(second_smallest_loop(list4))  # Output: None
  • Initializes smallest and second_smallest_val to infinity.
  • The loop updates these variables based on comparisons, ensuring second_smallest_val holds the second unique minimum found.

Method 4: Remove Minimum and Find New Minimum (Use with Caution)

This involves finding the minimum, removing its first occurrence, and then finding the minimum of the remaining list. Caution: This modifies a copy of the list and behaves differently depending on duplicates.

def second_smallest_remove(data_list):
    """Finds second smallest by removing the min. CAUTION with duplicates."""
    if not data_list or len(data_list) < 2:
        return None

    list_copy = data_list.copy() # Work on a copy
    min_value = min(list_copy)   # Find absolute minimum

    try:
        list_copy.remove(min_value) # Remove ONLY the first occurrence
    except ValueError: # Should not happen if list is not empty
        return None

    return min(list_copy) if list_copy else None


list1 = [3, 1, 4, 1, 5, 9, 2, 6]     # remove 1 -> [3, 4, 1, 5, 9, 2, 6] -> min is 1
list2 = [1, 1, 1, 3, 5]              # remove 1 -> [1, 1, 3, 5] -> min is 1
list3 = [10, 5, 20, 15]              # remove 5 -> [10, 20, 15] -> min is 10

print(second_smallest_remove(list1)) # Output: 1
print(second_smallest_remove(list2)) # Output: 1
print(second_smallest_remove(list3)) # Output: 10
  • This method effectively finds the second element if the list were sorted (like heapq.nsmallest). It does not find the second unique value reliably because remove() only targets the first instance. It's generally less efficient and less clear than other methods.

Finding the Second Largest Number

The logic mirrors finding the second smallest, with adjustments for comparisons and indexing.

Method 1: Using set, sorted, and Indexing (Unique Values)

Finds the second unique largest value.

def second_largest_unique_sorted(data_list):
  """Finds the second unique largest number using set and sorted."""
  if not data_list: return None
  unique_elements = set(data_list)
  if len(unique_elements) < 2: return None
  return sorted(unique_elements)[-2]

list1 = [3, 1, 4, 1, 5, 9, 2, 6, 9]         # Unique sorted: [1, 2, 3, 4, 5, 6, 9] -> 2nd largest is 6
list2 = [10, 10, 8, 5]                      # Unique sorted: [5, 8, 10] -> 2nd largest is 8

print(second_largest_unique_sorted(list1))  # Output: 6
print(second_largest_unique_sorted(list2))  # Output: 8

Method 2: Using heapq.nlargest()

Finds the second largest element, considering duplicates similarly to sorting.

import heapq # Required import

def second_largest_heapq(data_list):
  """Finds the second largest number using heapq.nlargest()."""
  if not data_list or len(data_list) < 2:
      try:
          if len(data_list) == 1: return None
          largest_two = heapq.nlargest(2, data_list)
          return largest_two[1] if len(largest_two) >= 2 else None
      except IndexError:
          return None
  else:
      return heapq.nlargest(2, data_list)[1]

list1 = [3, 1, 4, 1, 5, 9, 2, 6, 9] # Largest two [9, 9] -> 2nd is 9
list2 = [10, 5, 20, 15]             # Largest two [20, 15] -> 2nd is 15

print(second_largest_heapq(list1))  # Output: 9
print(second_largest_heapq(list2))  # Output: 15

Method 3: Manual Iteration (Looping)

Finds the second unique largest value.

def second_largest_loop(data_list):
  """Finds the second unique largest number using a loop."""
  if not data_list or len(set(data_list)) < 2: return None

  largest = float('-inf')
  second_largest_val = float('-inf')

  for number in data_list:
    if number > largest:
      second_largest_val = largest
      largest = number
    elif number > second_largest_val and number != largest:
      second_largest_val = number

  return second_largest_val if second_largest_val != float('-inf') else None

list1 = [3, 1, 4, 1, 5, 9, 2, 6, 9] # -> 6
list2 = [10, 10, 8, 5]            # -> 8

print(f"{list1} -> Second unique largest (loop): {second_largest_loop(list1)}") # Output: 6
print(f"{list2} -> Second unique largest (loop): {second_largest_loop(list2)}") # Output: 8

Method 4: Remove Maximum and Find New Maximum (Use with Caution)

Similar caveats apply as with second_smallest_remove. It finds the second element positionally after removing the first max.

def second_largest_remove(data_list):
    """Finds second largest by removing the max. CAUTION with duplicates."""
    if not data_list or len(data_list) < 2: return None
    list_copy = data_list.copy()
    max_value = max(list_copy)
    try:
        list_copy.remove(max_value) # Remove first occurrence
    except ValueError: return None
    return max(list_copy) if list_copy else None

list1 = [3, 1, 9, 5, 9] # remove 9 -> [3, 1, 5, 9] -> max is 9
list2 = [10, 20, 5, 15] # remove 20 -> [10, 5, 15] -> max is 15

print(f"{list1} -> Second largest (remove): {second_largest_remove(list1)}") # Output: 9
print(f"{list2} -> Second largest (remove): {second_largest_remove(list2)}") # Output: 15

Handling Edge Cases (Empty Lists, Insufficient Elements)

  • Empty Lists: All methods should ideally return None or raise an error. The examples above mostly return None.
  • Lists with One Element: There's no second smallest/largest. Methods should return None or raise an error.
  • Lists with Only One Unique Element: Methods designed to find the second unique value (Sort+Set, Loop) should return None. Methods based on position (heapq, Remove+Find) might return the single unique value depending on implementation details. Ensure the chosen method's behavior aligns with your needs for these cases. Robust functions include explicit checks (len(set(l)) < 2 or len(l) < 2).

Choosing the Right Method

  • Second Unique Value Needed:
    • set + sorted + index ([1] or [-2]): Clear, concise, good readability. Performance depends on sorting unique elements.
    • Manual Iteration Loop: More complex logic but potentially more efficient (O(N) time) as it avoids sorting.
  • Second Value Positionally (Allowing Duplicates):
    • heapq.nsmallest(2)/nlargest(2) + index ([1]): Efficient (especially for large lists where N is much larger than 2), handles duplicates like sorting. Often a good choice.
    • Remove Min/Max + Find New Min/Max: Generally not recommended due to inefficiency and ambiguity with duplicates.
  • Readability vs. Performance: Sorting methods are often easier to read initially. Iteration and heapq can be more performant for large lists if you only need the second value.

Conclusion

Finding the second smallest or largest number in a Python list requires careful consideration of how duplicates should be treated.

  • For the second unique value, converting to a set then sorting, or using a careful manual iteration are good options.
  • For the second element positionally (as if sorted, duplicates included), heapq.nsmallest/nlargest is often the most efficient and clear method.

Always consider edge cases like empty lists or lists with insufficient unique elements and choose the method that best fits your specific requirements for handling duplicates and performance.

Navigation