Python Code Optimizations: Part 3

Introduction

Part I and Part II of this tip can be viewed from the respective links.

Using the Code

1. Looping over Dictionary Keys and Values

for k in d:
    print k, '-->', d[k]

In the above example, we loop over the keys and then lookup the value. It is not very fast because it will have to rehash every key and do a lookup on it.

Instead, do this to get both keys and values.

for k,v in d.items():
    print k, '-->', v

Here, we're using tuple unpacking. So when you need both keys and values, loop over them directly and hence no lookups are required. Also, items() makes a list in the memory, which is not required.

Pythonic Way: use iteritems()

for k,v in d.iteritems():
    print k, '-->', v

2. Construct a Dictionary from Two Lists

Use zip().

names = ['steve','bill','jeff']
org = ['apple', 'microsoft', 'amazon']

d = dict(zip(names, org))

Pythonic Way: use izip()

from itertools import izip
d = dict(izip(names, org))

Many would think that izip will make a tuple at each iteration, but what it does is after the dictionary has consumed the tuple, we loop back around to make the next tuple, we re-use the previous one, so it can build the dictionary using the same tuple over and over again.

3. Counting with Dictionaries

colours = ['blue','red','green','red','red', 'green']
d = {}
for colour in colours:
    if colour not in d:
        d[colour] = 0
    d[colour] += 1

Better Way: use get() method

d = {}
for colour in colours:
    d[colour] = d.get(colour, 0) + 1

Pythonic Way: use defaultdict or Counter

from collections import defaultdict
d = defaultdict(int)
for colour in colours:
    d[colour] += 1

Although in the above code, d will be an object of defaultdict, which will have to be converted to a dict, if you intend to use it as one!