CODESARRAY

Lesson 7, by Codesarray, , 4 min read

Python dictionaries: keys, values, get, items and counting

Store Python values by key in a dictionary, look them up, add new ones, loop over them with items(), avoid a KeyError with get(), and count words.

A Python dictionary stores values by key, so you ask for something by name instead of by position. By the end of this page you will make a dictionary, look up and add entries, avoid a KeyError with get(), loop over every pair with items(), count how often words appear, and know why a lookup stays fast however big the dictionary gets.

Keys and values

A dictionary holds pairs: a key and its value. prices = {"apple": 3, "pear": 5} has two entries. prices["apple"] gives 3: you look up by key, not by position. Assigning to a new key, as in prices["kiwi"] = 7, adds an entry. Assigning to a key that already exists replaces its value.

get, in and the KeyError

Ask for a key that is missing and Python raises an error: prices["fig"] gives KeyError: 'fig'. The safe way is get(). prices.get("fig", 0) returns the default, 0, instead of crashing. If you leave the default out, get() returns None. The in test checks keys: "pear" in prices is True, and 3 in prices is False, because 3 is a value, not a key.

Looping with items

Call items() to get each key and value together. for name, price in prices.items() gives you every pair, in the order you added them. A key can be a string, a number or a tuple, but not a list, because a key must never change: {[1, 2]: 5} raises TypeError: cannot use 'list' as a dict key (unhashable type: 'list').

prices = {"apple": 3, "pear": 5}
print(prices["apple"])
prices["kiwi"] = 7
print(len(prices))
print(prices.get("fig", 0))

for name, price in prices.items():
    print(name, price)

words = ["py", "go", "py"]
counts = {}
for w in words:
    counts[w] = counts.get(w, 0) + 1
print(counts)
3
3
0
apple 3
pear 5
kiwi 7
{'py': 2, 'go': 1}

Line 1 builds the dictionary and line 2 looks up "apple". Line 3 adds "kiwi", so line 4 prints 3 entries. Line 5 asks for "fig" with a default of 0, and prints 0. Lines 7 and 8 loop over every pair. Lines 10 to 14 count words: counts starts empty, and for each word counts.get(w, 0) + 1 starts a new word at zero and adds one. The result is py twice and go once.

Why lookup is fast

A dictionary turns the key into a number, called a hash, and jumps straight to that spot. So finding a key is O(1), even with a million entries. Looking for an item in a list checks the items one by one, which is O(n). That is also why keys must never change: if a key changed, its hash would point to the wrong spot.

What you learned

  • A dictionary maps keys to values, and prices["apple"] looks up by key.
  • Assigning to a new key adds an entry; assigning to an existing key replaces its value.
  • get(key, default) returns the default instead of raising a KeyError, and in checks keys.
  • items() loops over keys and values together, in the order they were added.
  • Lookup is O(1) because of hashing, and a list cannot be a key.

Common questions

How do I make an empty dictionary? Use {} or dict(). Both give {}. Remember that an empty {} is a dictionary, not a set.

How do I remove a key? Use del prices["pear"], or prices.pop("pear"), which also hands back the value. Both raise a KeyError if the key is missing.

Are dictionaries ordered? Yes. Since Python 3.7, a dictionary keeps the order in which you added its keys, so list(prices) is ['apple', 'pear', 'kiwi'].

Is there a shorter way to count? Yes. Counter("banana") from collections gives Counter({'a': 3, 'n': 2, 'b': 1}). The get() loop above shows what it does underneath.

Can two keys be the same? No. A second assignment to the same key replaces the first value. Even 1 and 1.0 count as the same key, because they are equal.

This lesson follows on from Python tuples and sets. Next in the series: functions.

This article goes with the video Dictionaries. Watch it on YouTube.