CODEWITHSUNDEEP
pythonpython-3.xdictionarygraph-theoryattributeerror
AdKor
AdKor

Reputation: 255

How can I correct the error ' AttributeError: 'dict_keys' object has no attribute 'remove' '?

I'm trying to code the shortest path finder using Dijkstra's algorithm but it doesn't seem to be working. I can't figure out what the problem is. I'm working on Python 3.5 and following this video.

Here is the code:

graph = {
    'A': {'B': 10, 'D': 4, 'F': 10},
    'B': {'E': 5, 'J': 10, 'I': 17},
    'C': {'A': 4, 'D': 10, 'E': 16},
    'D': {'F': 12, 'G': 21},
    'E': {'G': 4},
    'F': {'E': 3},
    'G': {'J': 3},
    'H': {'G': 3, 'J': 3},
    'I': {},
    'J': {'I': 8},
}

def dijkstra(graph, start, end):
    D = {}
    P = {}
    for node in graph.keys():
        D[node]= -1
        P[node]=""
    D[start]=0
    unseen_nodes=graph.keys()
    while len(unseen_nodes) > 0:
        shortest=None
        node=' '
        for temp_node in unseen_nodes:
            if shortest==None:
                shortest = D[temp_node]
                node = temp_node
            elif D[temp_node]<shortest:
                    shortest=D[temp_node]
                    node=temp_node
        unseen_nodes.remove(node)
        for child_node, child_value in graph[node].items():
            if D[child_node] < D[node] + child_value:
                D[child_node] = D[node] + child_value
                P[child_node]=node
    path = []
    node = end
    while not (node==start):
        if path.count(node)==0:
            path.insert(0, node)
            node=P[node]
        else:
            break
    path.insert(0, start)
    return path

and here is the error message:

AttributeError: 'dict_keys' object has no attribute 'remove'

Upvotes: 25

Views: 48295

Answers (2)

cottontail
cottontail

Reputation: 23381

In Python 2, graph.keys() returns a list which defines a remove() method (see a demo here). graph.keys() being a list also means that it is a new copy of the graph's keys in its current state). In Python 3, it returns a dict_keys object which is a view of the dictionary's keys (which means whenever graph is modified, the view changes as well).

Since OP wants to create a new copy of the dict's keys that defines a remove() method, an alternative is to create a set. In other words, change

unseen_nodes = graph.keys()

to

unseen_nodes = set(graph)

Then a node can be removed using a remove method. An example would work as follows.

graph = {'a': 2, 'b': 1}
unseen_nodes = set(graph)
unseen_nodes.remove('a')   # remove node `a`
unseen_nodes               # the remaining nodes: {'b'}

An advantage of set over list is that it is much faster. For example, it is much faster to remove items from a set than from a list. The test below shows that removing items from a set is over 1000 times faster than removing items from a list.

from timeit import timeit
setup = '''
import random
lst = list(range(100000))
st = set(range(100000))
keys = iter(random.sample(range(100000), 100000))
'''

t1 = timeit('lst.remove(next(keys))', setup, number=100000)
t2 = timeit('st.remove(next(keys))', setup, number=100000)
print(t1/t2)     # 1330.4911104284974

Upvotes: 0

falsetru
falsetru

Reputation: 369424

In Python 3, dict.keys() returns a dict_keys object (a view of the dictionary) which does not have remove method; unlike Python 2, where dict.keys() returns a list object.

>>> graph = {'a': []}
>>> keys = graph.keys()
>>> keys
dict_keys(['a'])
>>> keys.remove('a')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'dict_keys' object has no attribute 'remove'

You can use list(..) to get a keys list:

>>> keys = list(graph)
>>> keys
['a']
>>> keys.remove('a')
>>> keys
[]

unseen_nodes = graph.keys()

to

unseen_nodes = list(graph)

Upvotes: 48

Related Questions