You now have four containers. Picking the right one makes code shorter, faster, and clearer — picking the wrong one makes you write loops you didn't need.

The comparison

| | List | Tuple | Dictionary | Set | | --- | --- | --- | --- | --- | | Syntax | [1, 2] | (1, 2) | {"a": 1} | {1, 2} | | Ordered | yes | yes | yes (insertion) | no | | Changeable | yes | no | yes | yes | | Duplicates | yes | yes | keys unique | no | | Look up by | index | index | key | — |

The decision, in four questions

1. Do you look things up by name? Use a dictionary. Storing a user's email, age, and city? user["email"] beats remembering user[2] every time.

2. Do you only care whether something is present, with no duplicates? Use a set. Tags, allowed usernames, "which pages has this visitor seen" — all sets.

3. Should it never change? Use a tuple. A coordinate, an RGB colour, a fixed pair of values returned together.

4. Otherwise? Use a list. Ordered, changeable, allows duplicates — the sensible default when nothing above applies.

Picking wrong: a worked example

Say you want to count how many times each word appears. With a list:

words = ["a", "b", "a", "c", "b", "a"]
unique_words = []
counts = []

for word in words:
    if word in unique_words:
        counts[unique_words.index(word)] += 1
    else:
        unique_words.append(word)
        counts.append(1)

print(unique_words)
print(counts)

Two parallel lists you have to keep in sync, plus an .index() scan on every word. Now with a dictionary:

words = ["a", "b", "a", "c", "b", "a"]
counts = {}

for word in words:
    counts[word] = counts.get(word, 0) + 1

print(counts)

Three lines, nothing to keep in sync, and faster. The data structure did the work.

Combining them

Most real programs use several together. A list of dictionaries for records, a set for fast membership checks, tuples for fixed pairs:

people = [
    {"name": "Ada", "tags": {"python", "math"}},
    {"name": "Grace", "tags": {"compilers"}},
]

for person in people:
    if "python" in person["tags"]:
        print(person["name"])

Try it yourself

Count the words using a dictionary, then print the count for "a" and the number of unique words.

Checkpoint

• Look up by name → dictionary. Uniqueness or membership → set.
• Must not change → tuple. Anything else → list.
• The right structure removes loops rather than adding them.
• Real programs nest and combine all four.