Discussion 8: Linked Lists
Switch to Pensieve:
- Everyone: Go to pensieve.co, log in with your @berkeley.edu email, and enter your group number (which was in the email that assigned you to this lab).
Once you're on Pensieve, you don't need to return to this page; Pensieve has all the same content (but more features). If for some reason Penseive doesn't work, return to this page and continue with the discussion.
Getting Started
To get help from a TA, If you do not have an in-person TA, you can reach your TA using this Zoom link.
If there are fewer than 3 people in your group, feel free to merge your group with another group in the room.
Everybody say your name and your birthday and then tell the group about your favorite birthday party you've attended (either for your birthday or someone else's).
Pro tip: Groups tend not to ask for help unless they've been stuck for a looooooong time. Try asking for help sooner. We're pretty helpful! You might learn something.
Linked Lists
A linked list is a Link
object or Link.empty
.
You can mutate a Link
object s
in two ways:
- Change the first element with
s.first = ...
- Change the rest of the elements with
s.rest = ...
You can make a new Link
object by calling Link
:
Link(4)
makes a linked list of length 1 containing 4.Link(4, s)
makes a linked list that starts with 4 followed by the elements of linked lists
.
class Link:
"""A linked list is either a Link object or Link.empty
>>> s = Link(3, Link(4, Link(5)))
>>> s.rest
Link(4, Link(5))
>>> s.rest.rest.rest is Link.empty
True
>>> s.rest.first * 2
8
>>> print(s)
<3 4 5>
"""
empty = ()
def __init__(self, first, rest=empty):
assert rest is Link.empty or isinstance(rest, Link)
self.first = first
self.rest = rest
def __repr__(self):
if self.rest:
rest_repr = ', ' + repr(self.rest)
else:
rest_repr = ''
return 'Link(' + repr(self.first) + rest_repr + ')'
def __str__(self):
string = '<'
while self.rest is not Link.empty:
string += str(self.first) + ' '
self = self.rest
return string + str(self.first) + '>'
Drawing time: Pick a way for your group to draw diagrams. Paper, a whiteboard, or a tablet, are all fine. If you don't have anything like that, ask another group in the room if they have extra paper.
Q1: Strange Loop
In lab, there was a Link
object with a cycle that represented an infinite repeating list of 1's.
>>> ones = Link(1)
>>> ones.rest = ones
>>> [ones.first, ones.rest.first, ones.rest.rest.first, ones.rest.rest.rest.first]
[1, 1, 1, 1]
>>> ones.rest is ones
True
Implement strange_loop
, which takes no arguments and returns a Link
object
s
for which s.rest.first.rest
is s
.
Draw a picture of the linked list you want to create, then write code to create it.
s.rest.first.rest
to exist at all, the second element of s
, called
s.rest.first
, must itself be a linked list.
s
= Link(6, Link(Link(1)))
, then change s.rest.first.rest
to create the cycle.
Q2: Sum Two Ways
Implement both sum_rec
and sum_iter
. Each one takes a linked list of numbers
s
and a non-negative integer k
and returns the sum of the first k
elements
of s
. If there are fewer than k
elements in s
, all of them are summed. If
k
is 0 or s
is empty, the sum is 0.
Use recursion to implement sum_rec
. Don't use recursion to implement
sum_iter
; use a while
loop instead.
s.first
to the sum of the first k-1
elements in s.rest
. Your base case
condition should include s is Link.empty
so that you're checking whether s
is empty before ever evaluating s.first
or s.rest
.
total
, then repeatedly (in a while
loop) add
s.first
to total
, set s = s.rest
to advance through the linked list, and reduce k
by one.
Discussion time: When adding up numbers, the intermediate sums depend on the
order. (1 + 3) + 5
and 1 + (3 + 5)
both equal 9, but the first one makes 4
along the way while the second makes 8 along the way. For the same linked list
s
and length k
, will sum_rec
and sum_iter
both make the same
intermediate sums along the way?
For a summation, the order of additions doesn't affect the result, but for other operations this ordering matters. If you're not sure why, spend a few minutes talking to your TA about when it might make a difference.
Q3: Overlap
Implement overlap
, which takes two linked lists of numbers called s
and t
that are sorted in increasing order and have no repeated elements within each
list. It returns the count of how many numbers appear in both lists.
This can be done in linear time in the combined length of s
and t
by
always advancing forward in the linked list whose first element is smallest
until both first elements are equal (add one to the count and advance both) or
one list is empty (time to return). Here's a
lecture video clip
about this (but the video uses Python lists instead of linked lists).
Take a vote to decide whether to use recursion or iteration. Either way works (and the solutions are about the same complexity/difficulty).
Run in 61A Code if s is Link.empty or t is Link.empty:
return 0
if s.first == t.first:
return __________________
elif s.first < t.first:
return __________________
elif s.first > t.first:
return __________________
k = 0
while s is not Link.empty and t is not Link.empty:
if s.first == t.first:
__________________
elif s.first < t.first:
__________________
elif s.first > t.first:
__________________
return k
Document the Occasion
Please all fill out the attendance form (one submission per person per week).
Important: Please help put the furniture in the room back where you found it before you leave. Thanks!
Extra Challenge
This last question is similar in complexity to an A+ question on an exam. Feel free to skip it, but it's a fun one, so try it if you have time.
Q4: Decimal Expansion
Definition. The decimal expansion of a fraction n/d
with n < d
is an
infinite sequence of digits starting with the 0 before the decimal point and
followed by digits that represent the tenths, hundredths, and thousands place
(and so on) of the number n/d
. E.g., the decimal expansion of 2/3 is a zero
followed by an infinite sequence of 6's: 0.6666666....
Implement divide
, which takes positive integers n
and d
with n < d
. It
returns a linked list with a cycle containing the digits of the infinite decimal
expansion of n/d
. The provided display
function prints the first k
digits
after the decimal point.
For example, 1/22 would be represented as x
below:
>>> 1/22
0.045454545454545456
>>> x = Link(0, Link(0, Link(4, Link(5))))
>>> x.rest.rest.rest.rest = x.rest.rest
>>> display(x, 20)
0.04545454545454545454...
Run in 61A Code
while
statement:
>>> q, r = 10 * n // d, 10 * n % d
>>> tail.rest = Link(q)
>>> tail = tail.rest
>>> n = r
While constructing the decimal expansion, store the tail
for each n
in a
dictionary keyed by n
. When some n
appears a second time, instead of
constructing a new Link
, set its original link as the rest of the previous
link. That will form a cycle of the appropriate length.