Dictionary with keys in unicode

Question

Is it possible in Python to use Unicode characters as keys for a dictionary? I have Cyrillic words in Unicode that I used as keys. When trying to get a value by a key, I get the following traceback:

 Traceback (most recent call last):
 File "baseCreator.py", line 66, in <module>
    createStoresTable()
 File "baseCreator.py", line 54, in createStoresTable
    region_id = regions[region]
 KeyError: u'\u041c\u0438\u043d\u0441\u043a/\u041c\u043e\u0441\u043a\u043e\u0432\u0441\u043a\u0438\u0439\xa0'

Answer 1

Yes, it's possible. The error you're getting means that the key you're using doesn't exist in your dictionary.

To debug, try printing your dictionary; you'll see the repr of each key which should show what the actual key looks like.

Answer 2

Python 2.x converts both keys to bytestrings when comparing two keys for the purposes of testing whether a key already exists, accessing a value, or overwriting a value. A key can be stored as Unicode, but two distinct Unicode strings cannot both be used as keys if they reduce to identical bytestrings.

In []: d = {'a': 1, u'a': 2}
In []: d
Out[]: {'a': 2}

You can use Unicode keys, in some sense.

Unicode keys are retained in Unicode:

In []: d2 = {u'a': 1}
In []: d2
Out[]: {u'a': 1}

You can access the value with any Unicode string or bytestring that "equals" the existing key:

In []: d2[u'a']
Out[]: 1

In []: d2['a']
Out[]: 1

Using the key or anything that "equals" the key to write a new value will succeed and retain the existing key:

In []: d2['a'] = 5
In []: d2
Out[]: {u'a': 5}

Because comparing 'a' to an existing key was True, the value corresponding to that existing Unicode key was replaced with 5. In the initial example I give, the second key u'a' provided in the literal for d compares truthfully to the previously assigned key, so the bytestring 'a' was retained as the key but the value was overwritten with the 2.