How to Migrate Dictionary Python2 items() to Python3 items()

Learn how to replace Python 2's iteritems() with Python 3's memory-efficient items() view in dictionary operations for seamless code migration.

When migrating from Python 2 to Python 3, iteritems() removal is among the most common causes of breakage. Understanding the difference helps you update legacy code correctly.

The Python 3 Standard

In Python 3, iteritems() no longer exists. The items() method now returns a view object that's memory-efficient and dynamic: combining the best of both Python 2 methods.

data = {"a": 1, "b": 2, "c": 3}

items_view = data.items()
print(type(items_view))

for key, value in data.items():
    print(f"{key}: {value}")

Output:

<class 'dict_items'>
a: 1
b: 2
c: 3

View objects reflect dictionary changes in real-time:

data = {"x": 10}
view = data.items()

print(list(view))

data["y"] = 20
print(list(view))

Output:

[('x', 10)]
[('x', 10), ('y', 20)]

The Python 2 Legacy Behavior

Python 2 offered two distinct methods with different memory characteristics:


data = {"a": 1, "b": 2}

pairs_list = data.items()      # [('a', 1), ('b', 2)]

pairs_iter = data.iteritems()  # <dictionary-itemiterator>

For large dictionaries, items() in Python 2 consumed significant memory since it built an entire list of tuples.

Migrating Your Code

Replace all occurrences of iteritems() with items():

for key, value in config.iteritems():
    process(key, value)

for key, value in config.items():
    process(key, value)

Example with actual code:

config = {"host": "localhost", "port": 8080, "debug": True}

for key, value in config.items():
    print(f"{key} = {value}")

Output:

host = localhost
port = 8080
debug = True

The same applies to keys() and values():

for k in my_dict.iterkeys():
    pass
for v in my_dict.itervalues():
    pass

for k in my_dict.keys():
    pass
for v in my_dict.values():
    pass

Example:

data = {"x": 100, "y": 200, "z": 300}

print("Keys:", list(data.keys()))
print("Values:", list(data.values()))

Output:

Keys: ['x', 'y', 'z']
Values: [100, 200, 300]

Example:

data = {"a": 1, "b": 2, "c": 3}
items_list = list(data.items())

print(items_list[0])
print(items_list[1])
print(len(items_list))

Output:

('a', 1)
('b', 2)
3

Writing Compatible Code

For codebases that must support both Python versions during transition, use a compatibility approach:

import sys

if sys.version_info[0] >= 3:
    iteritems = lambda d: d.items()
else:
    iteritems = lambda d: d.iteritems()

for key, value in iteritems(my_dict):
    process(key, value)

Or use the six compatibility library:

import six

for key, value in six.iteritems(my_dict):
    process(key, value)

Example with six:

import six

data = {"name": "Alice", "age": 30, "city": "NYC"}

for key, value in six.iteritems(data):
    print(f"{key}: {value}")

Output (Python 3):

name: Alice
age: 30
city: NYC

Method Comparison

Method	Python 2 Behavior	Python 3 Behavior
`items()`	Returns list (heavy memory)	Returns view (lightweight)
`iteritems()`	Returns iterator (lightweight)	❌ Removed
`keys()`	Returns list	Returns view
`iterkeys()`	Returns iterator	❌ Removed
`values()`	Returns list	Returns view
`itervalues()`	Returns iterator	❌ Removed

Summary

Forget iteritems() entirely for modern Python development. In Python 3, items() provides memory-efficient view objects that serve the same purpose. Simply replace iteritems() with items() during migration: no other changes needed.