密钥检查中“key in dict”和“dict.get(key)”之间的区别

Question

When I was doing Leetcode 820 , building a Trie, there is a bug in my code. I found it, also corrected it, but did NOT understand why. Could anyone help?

I made a simple test code here. Program 1 and Program 2 are doing same thing, that is, given a word list words , it builds a Trie in reversed order of these words. Variable trie stored the root of Trie, variable leaves stored beginning character of each word. The only difference between program 1 and program 2 is in line # difference , and I focus on the different results of leaves in 2 programs.

words = ["time", "atime", "btime"]

# program 1
trie = dict()
leaves = []
for word in words:
    node = trie
    for c in reversed(word):
        if c not in node:  # difference
            node[c] = dict()
        node = node[c]
    leaves.append(node)
print("from program 1: ")
print(leaves)

# program 2
trie = dict()
leaves = []
for word in words:
    node = trie
    for c in reversed(word):
        if not node.get(c):  # difference
            node[c] = dict()
        node = node[c]
    leaves.append(node)
print("from program 2: ")
print(leaves)

However the outputs of leaves are totally different.

from program 1: 
[{'a': {}, 'b': {}}, {}, {}]
from program 2: 
[{}, {}, {}]

Could anyone help explain why results of leaves in program 1 and program 2 are different? Thanks so much in advance!

Answer 1

c not in node can only be True when node does not contain the key c .

Lots of things can make not node.get(c) return True. Any falsy value, for example, like an empty list , dict or str , or something like 0 or 0.0 , evaluates to boolean False, just like None does. So you are liable to get a false negative any time node contains c , but has a falsy value.

If you're absolutely sure that your mapping can not contain None values, you can do the following:

if node.get(c) is not None:

If course explicitly checking for a key is always better than implicitly checking is by other means because of unintended side-effects like that.

Answer 2

One obvious difference:

• key in dict will only give you a false value if the key is not there.
• dict.get(key) will give you the actual value for that key which is then treated as a truthy or falsey value. So, if the value for an existing key is falsey (such as zero or the empty string), the if statement will not fire (though it will in the previous bullet point).

The following example shows this:

>>> d = {1: 0, 2: 42}

>>> 1 in d
True

>>> if d.get(1): print("yes")
...

>>> 2 in d
True

>>> if d.get(2): print("yes")
...
yes

---

If, as it seems, you're just trying to establish whether a key exists, the key in dict method is the correct way to go (you don't really care what the value is, just whether the value exists).