Tuples and Sorting

Lists, strings, and tuples are all ordered sequences. However, unlike lists, a tuple is an immutable sequence: a sequence that cannot be changed after it has been created.

Tuples¶

Tuples are defined by the , (comma) operator. But commas are used in many places in Python. To resolve ambiguity, use () (parentheses) to indicate a sequence of comma-separated elements belong in a tuple, just like how [] (square brackets) indicate a sequence of comma-separated elements belong in a list.

w = (4, 7, 9)
x = ("once", "upon", "a", "time")
y = ()  # An empty tuple

# Parentheses can make things clearer but are often not needed:
z = "kevinl", "blue"

Retrieve elements from a tuple using standard sequence operations:

w = (4, 7, 9)
w[0], w[1], w[2]
len(w)

x = ("once", "upon", "a", "time")
x[4]

Practice: w, x, y, z¶

Given the tuples w = (4, 7, 9) and x = ("once", "upon", "a", "time"), which statement will result in an error?

y = w[2]
x[3][2]
z = ([1, 2], [3, 4])
w[1] = 35

Tuples in the wild¶

The dict.items() function returns a sequence of key-value pairs represented as tuples: (key, value).

fav_color = {"akap1204": "lavender", "ktgower": "pink", "aronacho": "fuchsia", "ksuhas16": "green"}
for uwnetid, color in fav_color.items():
    print(uwnetid, " prefers ", color)

Tuples are commonly used to return multiple values from a single function call. The returned tuple can then be unpacked into separate variables.

def read_data(fname):
    data = []
    label = []
    # ... code that populates the two lists ...
    return data, label

data, label = read_data("data/mnist.csv")

Reference semantics¶

Understanding mutability requires differentiating between variables (names) and the objects they refer to in memory.

Assigning a variable changes the variable’s value. It does not mutate (modify) the original object.

size = 6
lst2 = lst1

Mutating an object mutates (modifies) the object’s internal state, so any variable referencing that object will also see the change.

lst1[2] = 5
lst1.append(12)

Of the three ordered sequence types (lists, strings, and tuples), only lists are mutable.

Practice: Tracing memory¶

Trace the following code in Python Tutor. What will lst1, lst2, and lst3 evaluate to at the end of the execution?

lst1 = [7, 8, 9]
lst2 = ["e1", "e2"]

lst3 = lst2
lst2 = lst1

lst1[2] = 5
lst1.append(12)

Sorting¶

Python provides two ways to sort data: the sort() method and the sorted() built-in function.

• list.sort() is a method that sorts the given list in-place. It modifies the original list and returns None. It can only be used on lists.

• sorted(itr) is a built-in function that takes an iterable (like a list, string, tuple, or dictionary keys) and returns a new sorted list. It leaves the original iterable unchanged.

quote = "with great power there must also come great responsibility".split()

sorted(quote)
quote

quote.sort()
quote

Sometimes the default sorting behavior (alphabetical for strings, numerical for numbers) is not what you need. You can customize the sorting criteria using a sort key. This function is called on each element to determine the value that will be used for comparisons.

names = ["Isaac Newton", "Ada Lovelace", "Fig Newton", "Grace Hopper"]
sorted(names, key=len)

def last_name(name):
    return name.split()[1]

sorted(names, key=last_name)

itemgetter¶

When working with tuples or dictionaries, writing custom functions just to extract a specific index or key can be tedious. Python’s operator module provides itemgetter, a convenient tool for creating functions that extract specific elements. For example, consider the following students and their favorite numbers.

fav_number = [("Ann", 7), ("Raul", 6), ("Ted", 4), ("Lisa", 6)]

One way is to define a function that returns just the number in each pair and pass it as the sort key.

def get_second(pair):
    return pair[1]

sorted(fav_number, key=get_second)

Another way is to use itemgetter.

from operator import itemgetter

sorted(fav_number, key=itemgetter(1))

itemgetter can even index into dictionaries too!

fav_number = [
    {"name": "Robert", "number": 8},
    {"name": "Alice", "number": 9},
    {"name": "Tina", "number": 7}
]

sorted(fav_number, key=itemgetter("number"))

Multiple sorting criteria¶

To sort by multiple criteria (e.g., sort by number, and sort by name only if there’s a tie), pass multiple arguments to itemgetter. Order matters: put the most important criterion first!

fav_number = [("Robert", 8), ("Alice", 9), ("Tina", 10), ("James", 8)]

sorted(fav_number, key=itemgetter(1, 0))

Since Python preserves the order of ties, an alternative approach is to sort by the least important criterion first before proceeding to the most important criterion last:

sorted_by_name = sorted(fav_number, key=itemgetter(0))
sorted_by_number_then_name = sorted(sorted_by_name, key=itemgetter(1))