1. High level version Partition AlmostQuickSort
9/21/2018
CS 303 – Quick Sort
Lecture 16

2. QuickSort QuickSort(l,r) If (r-l) < 1 Return i  Partition(l,r)
QuickSort(l,i-1)
QuickSort(i+1,r)
Partition(l,r)
<= x x
>=x r l i
9/21/2018
CS 303 – Quick Sort
Lecture 16

3. Partition How can you Partition, so that: Partition is O(n) Why?
If both these conditions are met, then QuickSort is O(n log n)
If not...not!
Discuss various Partition techniques
scan from l,r inwards – swap when stuck
scan from l  r – swap when <x
Discuss choice of pivot
use statistical sample to improve odds
9/21/2018
CS 303 – Quick Sort
Lecture 16

4. AlmostQuickSort AlmostQuickSort(l,r) If (r-l) < 15 Return
i  Partition(l,r)
AlmostQuickSort(l,i-1)
AlmostQuickSort(i+1,r)
QuickSort(l,r)
InsertionSort(l,r)
Partition becomes easier to code
Final InsertionSort is O(n)
alas...(Almost)QuickSort is still O(n2) (in theory...but...)
9/21/2018
CS 303 – Quick Sort
Lecture 16