Kotlin

Created By: Geoffrey Challen

/ Updated: `2022-11-01`

We continue our discussion of sorting algorithms by introducing the wild child: Quicksort.
Quicksort *can* achieve best-case sorting behavior while using less space than Mergesort.
**But**, Quicksort also has some pathological cases we need to understand.
Let's get started!

You knew it was coming...

Mergesort was our first recursive sorting algorithm. It employed a bottom-up approach—first breaking the array into individual chunks, and then merging them back together.

Quicksort is another recursive approach, but it works differently. Let's first examine its operation at a high level, and then break it down further.

Like Mergesort, Quicksort is also based on another building block: partitioning. Let's see how that works:

You'll get to complete `partition`

on our next homework!
But we can at least experiment with it using the method built-in to our playground:

Next, first we'll implement Quicksort. Then we'll analyze its performance!

Next, let's build a recursive sorting algorithm based on `Partitioner.partition`

!
Once we have a partition method, completing the implementation is quite straightforward!

Next, let's use the Quicksort implementation we completed above to experiment with its performance. Surprises are in store!

What's going on here? Let's consider things visually.

Our next quiz will focus on sorting algorithms, and particularly the ones that we've covered together:

Let's review salient aspects of sorting performance. Specifically: best and worst case inputs and runtime, and memory usage.

Sorting algorithms represent a fascinating set of tradeoffs between different performance attributes. For example:

**Have a very small array to sort?**Insertion sort can actually outperform other algorithms on small arrays, because the recursive calls needed by Mergesort and Quicksort have some overhead associated with them.**Want predictable performance?**Mergesort has your back. O(n log n), O(n log n), O(n log n). Sometimes it's actually more important that a sort take a predictable amount of time than that it be completely optimal.**Short on space?**Quicksort's space utilization is substantially smaller than Mergesort, and with good pivot selection it can achieve similar performance.

Timsort, the default sorting algorithm used by several languages including Python and Java, actually combines elements of both insertion sort and Mergesort, in addition to some other tricks.

As a final note, let's discuss sort algorithm *stability*.
Stability is a desirable property of sorting algorithms.
It means that items with equal values will not change positions while the array is sorted.

Why is stability desirable? Because it allows us to run a sorting algorithm multiple times on complex data and produce meaningful results.

For example, imagine that want a list of restaurants sorted first by cuisine and then by name.
Assuming our sorting algorithm is *stable*, we can accomplish this by first sorting the list by name and then by cuisine.
However, if the sorting algorithm in *unstable* that second sort by cuisine will destroy the results of the sort by name, and render the combination meaningless.

Need more practice? Head over to the practice page.