Joannah Nanjekye [Joannah Nanjekye] - Python 2 and 3 Compatibility: With Six and Python-Future Libraries

In Python 3, dict.keys() does not return a list like in Python 2. It instead returns a view object. The returned dict_keys object is more like a set than a list. It actually returns iterable keys of a given dictionary.

There is a dict.iterkeys() method in Python 2 to return iterable dictionary keys and this method works the same way dict.keys() works. To get iterable keys in both Python 2 and 3, we use six.iterkeys .

The six.iterkeys() method replaces dictionary.iterkeys() in Python 2 and dictionary.keys() in Python 3.

There is also another neutral way of getting iterable keys by referring to the dictionary itself. For example, this Python 2 code:

is the same as this:

This works in both Python 2 and 3.

In summary, for neutral compatibility for iterable dictionary keys:

• Use six.iterkeys() , which replaces dictionary.keys() in Python 3 and dictionary.iterkeys() in Python 2.

• You can also refer to the dictionary to get iterable keys.

In Python 2, if we want to work with an iterable of dictionary values, then we write code that looks like this:

We call itervalues() in the dictionary. In contrast, in Python 3 we have to call values() in the dictionary to achieve the same.

To support both Python 2 and 3, we have to find a middle ground using one of these options. Use:

• itervalues() from six

• itervalues() from python-futures builtins module

Similarly, dictionary.items() returns a list of dictionary items in Python 2 and dict_item view objects. To get iterable items in Python 2, we use dict.iteritems() .

While in Python 3, we use dict.items() .

There are a couple of ways we can achieve compatibility. future and six have wrappers for this.

In Python 2, we simply use keys() to return a list of keys from a dictionary.