Skip to content
SimpleScope

Sorting

Insertion sort vs Counting sort

Insertion sort and Counting sort both put an array in order, in different ways. Insertion sort: insert each value into a sorted prefix. Counting sort: count each value, turn the counts into positions and place the values, with no comparisons. Counting sort compares no values: it takes O(n + k), where k is the number of possible values (or digit values) and d the number of digits (radix sort), while insertion sort takes O(n²) in the worst case. Counting sort is faster only when k and d are small compared with n; insertion sort makes no assumption about the values.

At a glance

PropertyInsertion sortCounting sort
TechniqueInsert each value into a sorted prefixCount each value, turn the counts into positions and place the values, with no comparisons
Compares values with each otherYesNo
Best case timeO(n)O(n + k)
Average case timeO(n²)O(n + k)
Worst case timeO(n²)O(n + k)
Extra spaceO(1)O(n + k)
Stable (equal values keep their order)YesYes
In place (no extra array)YesNo
Needs sorted inputNoNo

When to choose each

Choose insertion sort when

the array is small or nearly sorted.

Avoid it when

the array is large and in random order.

Choose counting sort when

the values are whole numbers in a range k that is small compared with n.

Avoid it when

the range of values is large, or the values are not whole numbers.

The same inputs, counted

These numbers come from running both real implementations: operations (comparisons plus writes, a swap counting as two writes). Counting sort makes no comparisons, so its count is writes only. It includes the k + 1 cells of the count array.

InputInsertion sortCounting sort
Mixed3047
Already sorted1247
Reversed4847
Nearly sorted1447
Duplicates3647

See them run

More comparisons