Everything I know about Python...

Learn to Write Pythonic Code!

Check out the book Writing Idiomatic Python!

Discuss Posts With Other Readers at discourse.jeffknupp.com!

Python Dictionaries

Aside: one thing I dislike about the official Python documentation is that only a small percentage of entries have example code. We should change that...)

One of the keys to becoming a better Python programmer is to have a solid grasp of Python's built-in data structures. Using the structured format below, today you'll learn what a dict is, when to use it, and see example code of all of its member functions. I have some other data structures in the works, so this may turn into a little series.

Dictionary

AKA

"Associate Array", "Map", "Hash Map", "Unordered Map"

Library

built-in

Description

Contains a series of key -> value mappings where the "key" is of any type that is hashable (meaning it has both a __eq__() and a __hash__() method). The "value" may be of any type and value types need not be homogeneous.

That means, for example, we can have a dictionary where some keys map to strings and others to ints. Probably not a great idea in practice, but there's nothing stopping you from doing it.

What Makes it Special

The conceptual implementation is that of a hash table, so checks for existence are quite fast. That means we can determine if a specific key is present in the dictionary without needing to examine every element (which gets slower as the dictionary gets bigger). The Python interpreter can just go to the location key "should be" at (if it's in the dictionary) and see if key is actually there.

Construction

Literal

  • {}: pair of braces for empty dictionary
  • {1:2, 3:4}: comma-separated list of the form key: value enclosed by braces

    Constructor

  • dict(one=2, three=4): using dict() with keyword arguments mapping keys to values (where one and two are valid identifiers)

  • dict([(1, 2), (3, 4)]): using dict() with an iterable containing iterables with exactly two objects, the key and value
  • dict(zip([1, 3], [2, 4])): using dict() with two iterables of equal length; the first contains a list of keys and the second contains their associated values.
  • dict({1:2, 3:4}): using dict() with the literal form as an argument. This is silly. Why would you want this?

Mutability

mutable

Ordering

undefined

When to Use It

When describing what you want to do, if you use the word "map" (or "match"), chances are good you need a dictionary. Use whenever a mapping from a key to a value is required.

Example Usage

state_capitals={
    'New York': 'Albany',
    'New Jersey': 'Trenton',
    }

"New York" is a key and "Albany" is a value. This allows us to retrieve a state's capital if we have the state's name by doing capital = state_capitals[state]

How Not to Use It

Remember, the great thing about dictionaries is we can find a value instantly, without needing to search through the whole dictionary manually, using the form value = my_dict['key'] or value = my_dict.get('key', None).

If you're searching for a value in a dictionary and you use a for loop, you're doing it wrong. Stop, go back, and read the previous statement.

All too often in beginner code I see the equivalent of the following (continuing the previous example):

state_im_looking_for = 'New Jersey'
my_capital = ''

for state in state_capitals:
    if state == state_im_looking_for:
        my_capital = state_capitals[state]

Or like this:

state_im_looking_for = 'New Jersey'
my_capital = ''

for state, capital in state_capitals.items():
    if state == state_im_looking_for:
        my_capital = capital

Methods and Uses

d.clear()

Remove all entries in d

Returns

N/A

Raises

N/A

Examples

Delete all items in a dictionary

d.clear()

d.copy()

Make a shallow copy of d. The dictionary returned by d.copy() will have the same references as d, not copies of the items.

Returns

A new dict, representing a shallow copy of d

Raises

N/A

Examples

Create copy of a dictionary

d = {1: 'a', 2: 'b', 3: 'c'}
copied_dict = d.copy()
copied_dict # {1: 'a', 2: 'b', 3: 'c'}
d[1] = 'z'
copied_dict # {1: 'a', 2: 'b', 3: 'c'}

del k[d]

Used to remove a value from a dictionary

Returns

N/A

Raises

KeyError if key is not in dictionary

Examples

Delete entry with key 'hello'

my_dictionary = {'hello': 1, 'goodbye': 2}
del my_dictionary['hello']
print(my_dictionary)
# {'goodbye': 2}

dict.fromkeys(seq[, value])

Create a new dictionary with the same keys as seq. If value is provided, each item's value is set to value. If value is not set, all item values are set to None

Returns

N/A

Raises

N/A

Examples

Create a dictionary from a list with all values initialized to 0

my_list = [1, 2, 3]
my_dictionary = dict.fromkeys(my_list, 0)
my_dictionary # {1: 0, 2: 0, 3: 0}

Create a dictionary from a dictionary with all values automatically initialized to None

my_dictionary = {1: 1, 2: 2, 3: 3}
new_dictionary = dict.fromkeys(my_dictionary)
my_dictionary # {1: None, 2: None, 3: None}

d.get(key[, default)

Used to retrieve the value associated with key key. The value of default is returned if key is not in d (rather than raising a KeyError). The default value of default is None.

Returns

Roughly equivalent to:

def get(key, default=None):
    if key in d:
        return d[k]
    else:
        return default

Raises

N/A

Examples

Get a key's value or None if the key isn't present

{1: 'a', 2: 'b'}.get(3)

k in d

Used to iterate over the keys, values, or both of the dictionary.

Returns

N/A

Raises

N/A

Examples

Iterate over keys

for key in my_dictionary:

Iterate over (key, value) tuples

for key, value in my_dictionary.items():

Iterate over values

for value in my_dictionary.values():

Check for existence

haystack = {}
# ...
if 'needle' in haystack:

iter(d)

Used to iterate over the keys of d

Returns

An iterator which iterates over the keys of d

Raises

StopIteration when d has no more keys

Examples

Iterate over keys

for key in my_dictionary:

d[key]

Used to access the value corresponding to the key key in d.

Returns

Value associated with the key (heterogeneous)

Raises

KeyError when key is not a member of d.

Examples

capitals = {'New York': 'Albany'}`
capital_of_ny = capitals['New York']`
print capital_of_ny`
'Albany'

len(d)

Used to determine the number of entries in a dictionary

Returns

Length of dictionary d

Raises

N/A

Examples

print 'dictionary has {} entries'.format(len(d))

k not in d

Used for negative existence check. Equivalent to not key in value

Returns

True if key is not in value, False otherwise

Raises

N/A

Examples

Check for negative existence

haystack = {}
# ...
if 'needle' not in haystack:

d.keys()

Iterate over the keys in a dictionary

Returns

An iterable over all of the keys in d (in an unspecified order)

Raises

StopIteration when d has no more keys

Examples

Iterate over keys:

for key in d.keys():

d.values()

Iterate over the values in a dictionary

Returns

An iterable over all of the values in d (in an unspecified order)

Raises

StopIteration when d has no more values

Examples

Iterate over values:

for value in d.values():

d.items()

Iterate over the elements ((key, value) pairs) in a dictionary

Returns

An iterable over all of the (key, value) pairs in d (in an unspecified order). Each (key, value) pairs is represented as a tuple.

Raises

StopIteration when d has no more elements

Examples

Iterate over items:

for key, value in d.items():

Note that, in the example, we can use multiple assignment to assign key to the key and value to the value of each item directly in the for loop.

d.pop(key[, default])

Used to remove an item from a dictionary and return its associated value

Returns

d[key] if key is in d. If key is not in d but default is specified, the default value is returned instead.

Raises

KeyError if key is not in dictionary and no default is specified

Examples

Delete entry with key 'hello' and print its value

my_dictionary = {'hello': 1, 'goodbye': 2}
hello_value = my_dictionary.pop('hello')
print(hello_value)
# 1
print(my_dictionary)
# {'goodbye': 2}

With default specified

my_dictionary = {'hello': 1, 'goodbye': 2}
foo_value = my_dictionary.pop('foo', None)
print(foo_value)
# None
print(my_dictionary)
# {'goodbye': 2}

With no default specified

my_dictionary = {'hello': 1, 'goodbye': 2}
foo_value = my_dictionary.pop('foo')
# KeyError: 'foo'

d.popitem()

Pop (i.e. delete and return) a random element from the dictionary

Returns

A (key, value) tuple if d is not empty.

Raises

KeyError if d is empty. I personally think that's a stupid exception to raise since no key was ever specified, but, hey, I didn't write the language.

Examples

Destructively iterate over values:

try:
    key, value = d.popitem():
    print 'Got {}: {}'.format(key, value)
except KeyError:
    print 'Done'

d.setdefault(key[, default])

Get a key from the dictionary or, if it's not there, insert it with a default value and return that. default, erm, defaults to None

Returns

d[key] if key is in d.

If not, do d[key] = default and then return d[key] (which will always return default).

Raises

N/A

Examples

Count the number of times each word is seen in a file:

words = {}

for word in file:
    occurrences = words.setdefault(word, 0)
    words[word] = occurrences + 1

d.update(other)

Update a dictionary with the keys and values in other, overwriting existing keys and values if there is any overlap.

Returns

None

Raises

N/A

Examples

Merge two dictionaries:

first = {'a': 1}
second = {'b': 2}
first.update(second)
print first
# {'a': 1, 'b': 2}
print second
# {'b': 2}

Using keyword arguments for other:

first = {'a': 1}
first.update(b=2, c=3)
print first
# {'a': 1, 'c': 3, 'b': 2}
Posted on by

Discuss Posts With Other Readers at discourse.jeffknupp.com!

comments powered by Disqus
Web Analytics