# Shaker Sort

This algorithm (AKA cocktail sort) has a complexity of O(n²).

The procedure is similar to the bubble sort. In fact, the only difference is that it sorts in both directions.

```
def shaker_sort(l):
n = len(l)
swap_count = 1
while swap_count > 0:
swap_count = 0
# forward
for i in range(n - 1):
if l[i] > l[i + 1]:
l[i], l[i + 1] = l[i + 1], l[i]
swap_count += 1
# backwards
for i in range(n - 1, 0, -1):
if l[i] < l[i - 1]:
l[i], l[i - 1] = l[i - 1], l[i]
swap_count += 1
n -= 1
```

As you can see from the code above, for each while loop there is a forward pass and a backwards pass. This fact turns this algorithm more efficient than bubble sort.

```
l = [1, 32, 3, 45, 5, 6, 8, 0]
shaker_sort(l)
print l # [0, 1, 3, 5, 6, 8, 32, 45]
```