Source: this section is heavily based on Chapter 9 of [ThinkCS].
We saw earlier that we could group together pairs of values by surrounding with parentheses. Recall this example:
>>> year_born = ("Paris Hilton", 1981)
The pair is an example of a tuple. This is an example of a data structure --- a mechanism for grouping and organizing data to make it easier to use. Generalizing this, a tuple can be used to group any number of items into a single compound value. Syntactically, a tuple is a comma-separated sequence of values. Although it is not necessary, it is conventional to enclose tuples in parentheses:
>>> julia = ("Julia", "Roberts", 1967, "Duplicity", 2009, "Actress", "Atlanta, Georgia")
Tuples are useful for representing what other languages often call records --- some related information that belongs together, like your student record. There is no description of what each of these fields means, but we can guess. A tuple lets us "chunk" together related information and use it as a single thing.
Tuples are very similar to lists and support the same sequence operations as lists. The index operator selects an element from a tuple:
>>> julia[2] 1967
However, tuples are not the same as lists. Next to the different syntax (parentheses instead of square brackets), an important difference is that tuples are unmutable. If we try to use item assignment to modify one of the elements of the tuple, we get an error:
>>> julia[0] = "X" TypeError: 'tuple' object does not support item assignment
Similarly, we can not add elemnts to a tuple, or remove them; once Python has created a tuple in memory, it cannot be changed.
Of course, even if we can't modify the elements of a tuple, we can always make the julia variable reference a new tuple holding different information. To construct the new tuple, it is convenient that we can slice parts of the old tuple and join up the bits to make the new tuple. So if julia has a new recent film, we could change her variable to reference a new tuple that used some information from the old one:
>>> julia = julia[:3] + ("Eat Pray Love", 2010) + julia[5:] >>> julia ("Julia", "Roberts", 1967, "Eat Pray Love", 2010, "Actress", "Atlanta, Georgia")
To create a tuple with a single element (but you're probably not likely to do that too often), we have to include the final comma, because without the final comma, Python treats the (5) below as an integer in parentheses:
>>> tup = (5,) >>> type(tup) <class 'tuple'> >>> x = (5) >>> type(x) <class 'int'>
Given the similarity between Python's tuples and lists, a good question is why tuples exist at all. Why doesn't Python only have lists? There are two good reasons for the existence of tuples.
One is the performance of the code. Tuples are more efficient than lists: they consume less memory and the run time required to create them is smaller.
Why is this? We skimmed over this till now, but allocating memory in a computer is not a trivial operation; essentially, each time a program requires more memory, the operating system will have to search for a piece of memory that is still unused. This also applies to lists. If the operating system would have to look for a new piece memory each time an element is added to a list, adding elements to a list would be a rather slow operation. To avoid this, Python tries to be intelligent: it will anticipate the addition of elements in a list by reserving more memory than necessary at the moment of creation. The benefit is that adding elements is now faster. The side effect is however that lists will consume more memory than necessary. Python's tuples avoid this. As a tuple will never change, we know its memory consumption will never change. Hence, Python does not need to anticipate future additions to the tuple.
The second reason for having tuples relates to the readability of code written using tuples. Consider this piece of code:
julia = ("Julia", "Roberts", 1967, "Duplicity", 2009, "Actress", "Atlanta, Georgia") do_something ( julia ) print ( julia[0] )
For this piece of code, we can be sure that whatever the functionality of the function do_something is, at the end the string "Julia" will be printed. This makes it easy to understand what the third line of code is doing.
Consider now this piece of code:
julia = ["Julia", "Roberts", 1967, "Duplicity", 2009, "Actress", "Atlanta, Georgia"] do_something ( julia ) print ( julia[0] )
In this code, we can no longer be sure about what julia[0] will print. Consider this implementation of the do_something function:
def do_something ( l ): l[0] = "Hugh"
This function will change the list julia, and as a result the code will print "Hugh". Hence, to understand what the line print ( julia[0] ) does, we will need to check the documentation or source code of the function do_something. For tuples, this is not necessary: by using tuples, the programmer can communicate to another reader of the code that this data is not supposed to be changed. Indeed, any function that you will apply on this tuple, and that would try to change it, will yield an error message, hence making it easier to debug the code as well.
Python has a very powerful tuple assignment feature that allows a tuple of variables on the left of an assignment to be assigned values from a tuple on the right of the assignment. (We already saw this used for pairs, but it generalizes.)
(name, surname, b_year, movie, m_year, profession, b_place) = julia
This can also be shortened to
name, surname, b_year, movie, m_year, profession, b_place = julia
This does the equivalent of seven assignment statements, all on one easy line. One requirement is that the number of variables on the left must match the number of elements in the tuple.
One way to think of tuple assignment is as tuple packing/unpacking.
In tuple packing, the values on the left are 'packed' together in a tuple:
>>> b = ("Bob", 19, "CS") # tuple packing
In tuple unpacking, the values in a tuple on the right are 'unpacked' into the variables/names on the right:
>>> b = ("Bob", 19, "CS") >>> (name, age, studies) = b # tuple unpacking >>> name 'Bob' >>> age 19 >>> studies 'CS'
Once in a while, it is useful to swap the values of two variables. With conventional assignment statements, we have to use a temporary variable. For example, to swap a and b:
temp = a a = b b = temp
Tuple assignment solves this problem neatly:
a, b = b, a
The left side is a tuple of variables; the right side is a tuple of values. Each value is assigned to its respective variable. All the expressions on the right side are evaluated before any of the assignments. This feature makes tuple assignment quite versatile.
Naturally, the number of variables on the left and the number of values on the right have to be the same:
>>> (a, b, c, d) = (1, 2, 3) ValueError: need more than 3 values to unpack
Functions can always only return a single value, but by making that value a tuple, we can effectively group together as many values as we like, and return them together. This is very useful --- we often want to know some batsman's highest and lowest score, or we want to find the mean and the standard deviation, or we want to know the year, the month, and the day, or if we're doing some some ecological modelling we may want to know the number of rabbits and the number of wolves on an island at a given time.
For example, we could write a function that returns both the area and the circumference of a circle of radius r:
def f(r): """ Return (circumference, area) of a circle of radius r """ c = 2 * math.pi * r a = math.pi * r * r return (c, a)
We saw in an earlier chapter that we could make a list of pairs, and we had an example where one of the items in the tuple was itself a list:
students = [ ("John", ["CompSci", "Physics"]), ("Vusi", ["Maths", "CompSci", "Stats"]), ("Jess", ["CompSci", "Accounting", "Economics", "Management"]), ("Sarah", ["InfSys", "Accounting", "Economics", "CommLaw"]), ("Zuki", ["Sociology", "Economics", "Law", "Stats", "Music"])]
Tuples items can themselves be other tuples. For example, we could improve the information about our movie stars to hold the full date of birth rather than just the year, and we could have a list of some of her movies and dates that they were made, and so on:
julia_more_info = ( ("Julia", "Roberts"), (8, "October", 1967), "Actress", ("Atlanta", "Georgia"), [ ("Duplicity", 2009), ("Notting Hill", 1999), ("Pretty Woman", 1990), ("Erin Brockovich", 2000), ("Eat Pray Love", 2010), ("Mona Lisa Smile", 2003), ("Oceans Twelve", 2004) ])
Notice in this case that the tuple has just five elements --- but each of those in turn can be another tuple, a list, a string, or any other kind of Python value. This property is known as being heterogeneous, meaning that it can be composed of elements of different types.
We have already seen the following code:
xs = [1, 2, 3, 4, 5] for i in range(len(xs)): xs[i] = xs[i]**2
While correct, this type of list traversal is so common, that Python provides a nicer way to implement it:
xs = [1, 2, 3, 4, 5] for (i, val) in enumerate(xs): xs[i] = val**2
This code exploits lists-of-tuples: enumerate generates pairs of both (index, value) during the list traversal. Try this next example to see more clearly how enumerate works:
for (i, v) in enumerate(["banana", "apple", "pear", "lemon"]): print(i, v)0 banana 1 apple 2 pear 3 lemon
Another common type of program one may wish to write is the following:
xs = [1, 2, 3, 4, 5] ys = [3, 4, 5, 6, 7] for i in range(len(xs)): print (xs[i],ys[i])
Using the enumerate function we could rewrite this as:
xs = [1, 2, 3, 4, 5] ys = [3, 4, 5, 6, 7] for (i, val) in enumerate(xs): print (val,ys[i])
However, most programmers would not consider this to be a very clean solution. Python provides the zip function to write this code more elegantly:
xs = [1, 2, 3, 4, 5] ys = [3, 4, 5, 6, 7] for x, y in zip(xs,ys): print (x,y)
Like a zipper, the zip function combines elements of two given lists pairwise, and provides a list of the tuples that represent pairs from the two given list.
In combination with the enumerate function, one can now write code like the following:
xs = [1, 2, 3, 4, 5] ys = [3, 4, 5, 6, 7] for i, (x, y) in enumerate(zip(xs,ys)): xs[i] = x**2 ys[i] = y**2
Observe that in this code, the zip function generates pairs of elements from the xs and ys lists. The enumerate function subsequently adds the indexes of the pairs in this list.
- data structure
- An organization of data for the purpose of making it easier to use.
- immutable data value
- A data value which cannot be modified. Assignments to elements or slices (sub-parts) of immutable values cause a runtime error.
- mutable data value
- A data value which can be modified. The types of all mutable values are compound types. Lists and dictionaries are mutable; strings and tuples are not.
- tuple
- An immutable data value that contains related elements. Tuples are used to group together related data, such as a person's name, their age, and their gender.
- tuple assignment
- An assignment to all of the elements in a tuple using a single assignment statement. Tuple assignment occurs simultaneously rather than in sequence, making it useful for swapping values.
- heterogeneous list
- A list that contains elements of different types.
- generators
- Functions that will generate lists
| [ThinkCS] | How To Think Like a Computer Scientist --- Learning with Python 3 |