Python AttributeError Reference — 170+ “Object has no attribute X” Fixes (2026)

itsourcecode.com hosts 170+ documented fixes for Python AttributeError messages: the most common runtime error after TypeError. From beginner-level confusion ('NoneType' object has no attribute 'split') to library-version pain (module 'numpy' has no attribute 'float' after NumPy 1.20+), this archive groups every documented error by root cause. Each post explains why the attribute is missing and gives 2–4 alternative fixes with code examples.

What is an AttributeError?

A Python AttributeError is raised when you try to access an attribute or method on an object that doesn’t have it. The error message always tells you both the object type and the missing attribute — for example AttributeError: 'list' object has no attribute 'items' means you called .items() on a list, but only dicts have that method. The error has two forms:

• Instance form: 'X' object has no attribute 'Y' — the value at runtime isn’t the type you expected
• Module form: module 'X' has no attribute 'Y' — the function/class you’re importing no longer exists in that version of the library

How to debug any AttributeError in 4 steps

1. Read both names in the error. The message gives you the object type AND the missing attribute. Note both — they tell you what’s wrong together.
2. Print the type of the object. Add print(type(x)) on the line before the error. If the error says 'NoneType' object has no attribute 'split', the value is actually None — usually because a function returned nothing when you expected a string.
3. If the error is “module X has no attribute Y”, check the library version. Run pip show to see installed version. Compare to the version that had the attribute — often the API was deprecated and renamed. Common offenders: NumPy (np.float, np.int, np.long removed in 1.20+), pandas (.ix removed in 1.0+), Keras (keras.optimizers.Adam moved), Selenium (find_element_by_id removed in 4.0+).
4. Check the import path. from numpy import float and numpy.float are different namespaces and different lookups. Many AttributeErrors come from a sub-package that moved or was renamed across versions.

Most common root causes (in order of frequency)

• Value is None when you expected an object — function returned None, regex match failed, dict lookup returned None, BeautifulSoup didn’t find the element
• Library version mismatch — API was deprecated and removed (numpy 1.20+, pandas 1.0+, Keras 2.x → tf.keras, Selenium 4)
• Wrong type altogether — calling .items() on a list, .append() on a string, .read() on a dict
• Python 2 → 3 migration — old code using dict.iteritems(), dict.has_key(), urllib.urlopen
• Module re-org — imports from tensorflow.contrib (removed), keras.optimizers.Adam (moved to tf.keras.optimizers)
• Typo in attribute name — beginners hit this most, especially when an IDE doesn’t autocomplete

Featured AttributeError fixes by type

🟢 Python built-in types: NoneType, str, list, dict, set, tuple

These errors come from runtime type confusion — your code expects one type but got another. Almost always fixable with a one-line type check or rewriting the previous line that returned the wrong value.

NoneType errors (14 posts) — the most common AttributeError class. A function/method returned None when you expected a string, list, or object.

• ‘NoneType’ object has no attribute ‘split’
• ‘NoneType’ object has no attribute ‘strip’
• ‘NoneType’ object has no attribute ‘find’ (BeautifulSoup)
• ‘NoneType’ object has no attribute ‘text’
• ‘NoneType’ object has no attribute ‘cursor’ (database)

str errors (10 posts) — calling list/dict methods on a string.

• ‘str’ object has no attribute ‘keys’
• ‘str’ object has no attribute ‘get’
• ‘str’ object has no attribute ‘append’
• ‘str’ object has no attribute ‘remove’
• ‘str’ object has no attribute ‘write’
• ‘str’ object has no attribute ‘decode’ (Python 3 migration)

list errors (10 posts) — calling string/dict methods on a list.

• ‘list’ object has no attribute ‘items’
• ‘list’ object has no attribute ‘strip’ — 2026 Guide
• ‘list’ object has no attribute ‘shape’ — 2026 Guide
• ‘list’ object has no attribute ‘join’

dict errors (6 posts) — Python 2 methods removed in Python 3.

• ‘dict’ object has no attribute ‘iteritems’ (Py3 migration)
• ‘dict’ object has no attribute ‘has_key’ (Py3 migration)
• ‘dict’ object has no attribute ‘add’ — 2026 Guide
• ‘dict’ object has no attribute ‘read’

set, float, bytes errors

• ‘set’ object has no attribute ‘items’
• ‘float’ object has no attribute ‘split’

🐼 pandas (DataFrame, Series) — 13 posts

pandas had multiple API breaks between 0.x → 1.x → 2.x. Old code using .ix, .append(), or older indexing methods raises AttributeError on modern pandas.

• ‘DataFrame’ object has no attribute ‘ix’ (use .loc / .iloc)
• ‘DataFrame’ object has no attribute ‘unique’
• ‘DataFrame’ object has no attribute ‘concat’
• ‘DataFrame’ object has no attribute ‘write’
• ‘DataFrame’ object has no attribute ‘_get_object_id’
• ‘DataFrame’ object has no attribute ‘withColumn’ (PySpark vs pandas)
• ‘Series’ object has no attribute ‘strftime’
• ‘Series’ object has no attribute ‘split’
• Pandas .dt accessor with datetimelike values
• module ‘pandas’ has no attribute ‘read_csv’

🔢 NumPy (module + ndarray) — 16 posts

NumPy 1.20 deprecated and removed many type aliases (np.float, np.int, np.long, np.bool). Code from 2019 or earlier hits these errors immediately on modern NumPy.

• module ‘numpy’ has no attribute ‘object’
• module ‘numpy’ has no attribute ‘float’
• module ‘numpy’ has no attribute ‘long’
• module ‘numpy’ has no attribute ‘ndarray’ — 2026 Guide

🤖 TensorFlow & Keras — 15 posts

TensorFlow 2.x reorganized the API: tensorflow.contrib was removed, optimizers moved into tf.keras.optimizers, and many sub-modules were merged or deprecated.

• module ‘tensorflow.compat.v1’ has no attribute ‘contrib’
• module ‘tensorflow’ has no attribute ‘set_random_seed’
• module ‘keras.utils’ has no attribute ‘Sequence’
• module ‘keras.optimizers’ has no attribute ‘Adam’
• ‘Tensor’ object has no attribute ‘numpy’

🔥 PyTorch & TorchText — 3 posts

• module ‘torchtext.data’ has no attribute ‘Field’ (torchtext 0.9+)
• module ‘torch._C’ has no attribute ‘_cuda_setDevice’

📈 Matplotlib (module-level) — 2 posts

• module ‘matplotlib’ has no attribute ‘plot’ (use pyplot)
• module ‘matplotlib’ has no attribute ‘get_data_path’
• ‘AxesSubplot’ object has no attribute ‘savefig’

👁️ OpenCV (cv2) — 2 posts

• module ‘cv2’ has no attribute ‘xfeatures2d’ (need opencv-contrib)
• module ‘cv2’ has no attribute ‘_registerMatType’

🕷️ Selenium WebDriver — 3 posts

Selenium 4 removed the find_element_by_* methods. Use find_element(By.X, ...) instead.

• ‘WebDriver’ object has no attribute ‘find_element_by_class_name’
• module ‘selenium.webdriver’ has no attribute ‘PhantomJS’ (deprecated)
• module ‘selenium.webdriver’ has no attribute ‘Opera’

⚙️ Standard library & misc modules

• module ‘datetime’ has no attribute ‘now’
• module ‘asyncio’ has no attribute ‘run’ (Python 3.7+)
• module ‘collections’ has no attribute ‘Iterable’ (Py3.10+ moved to collections.abc)
• module ‘urllib’ has no attribute ‘urlopen’ (Py3: use urllib.request)
• AttributeError: __enter__ (context manager protocol)
• module ‘h11’ has no attribute ‘Event’
• ‘entrypoints’ object has no attribute ‘get’
• ‘API’ object has no attribute ‘search’ (Tweepy)

2026 Updated Guides — featured AttributeError fixes

These guides were rewritten or expanded in 2026 with current library versions, minimal reproductions, and 3–4 alternative fixes each.

• module ‘numpy’ has no attribute ‘ndarray’ — NumPy 1.20+ migration
• ‘list’ object has no attribute ‘strip’ — String method confusion
• ‘list’ object has no attribute ‘shape’ — NumPy vs list confusion
• ‘dict’ object has no attribute ‘add’ — set vs dict confusion
• module ‘pandas’ has no attribute ‘read_csv’ — shadowed import
• ‘DataFrame’ object has no attribute ‘ix’ — pandas 1.0+ migration (use .loc / .iloc)
• Pandas .dt accessor error — datetime column conversion
• ‘Tensor’ object has no attribute ‘numpy’ — TensorFlow eager execution
• ‘WebDriver’ has no attribute ‘find_element_by_class_name’ — Selenium 4 migration
• module ‘asyncio’ has no attribute ‘run’ — Python 3.7+ async
• ‘tensorflow.compat.v1’ has no attribute ‘contrib’ — TF 2.x migration
• ‘set’ object has no attribute ‘items’ — set vs dict
• ‘list’ object has no attribute ‘items’ — list of dicts vs dict
• AttributeError: __enter__ — using with on non-context-manager

Frequently asked questions

What’s the difference between AttributeError and TypeError?

AttributeError says “this object exists but doesn’t have the attribute/method you tried to access.” TypeError says “the value’s type is wrong for this operation entirely.” For example, None.split() raises AttributeError (NoneType has no .split); None + 5 raises TypeError (you can’t add NoneType to int). Both are runtime errors caused by Python’s dynamic typing, but they fail at different points.

How do I fix “‘NoneType’ object has no attribute X”?

The variable is None when you expected it to be an object. Most common causes: (1) A function returned nothing (no return statement, or fell off the end of an if). (2) dict.get(key) when the key isn’t present (returns None by default). (3) BeautifulSoup’s .find() returned None because the element doesn’t exist. (4) A database query returned no rows. Fix by adding a None check: if x is not None: x.do_thing() — or fix the upstream code that’s producing None.

Why does “module ‘numpy’ has no attribute ‘float'” / ‘int’ / ‘long’ happen?

NumPy 1.20 (released January 2021) deprecated np.float, np.int, np.long, np.bool, and np.object. NumPy 1.24 (Dec 2022) removed them entirely. Replace them with: np.float64, np.int64, Python’s built-in float/int/bool, or downgrade to NumPy < 1.20 if you can't change the code.

Why does “DataFrame object has no attribute ‘ix'” happen?

.ix was deprecated in pandas 0.20 and removed in 1.0 (January 2020). It was ambiguous because it accepted both labels and positions. Replace with .loc (label-based) or .iloc (integer position-based). For df.ix[5, 'col'], use df.loc[5, 'col'] if 5 is a row label, or df.iloc[5, df.columns.get_loc('col')] if 5 is a position.

Why does “object has no attribute ‘iteritems’ / ‘has_key'” happen?

You’re running Python 3 code that used Python 2 methods. In Python 3, dict.iteritems() became dict.items(), and dict.has_key(k) became k in dict. Run a Python 2 → 3 conversion with 2to3 tool, or manually update: for k, v in d.iteritems() → for k, v in d.items(); if d.has_key('x') → if 'x' in d.

Why does “WebDriver has no attribute find_element_by_class_name” happen?

Selenium 4 (October 2021) removed all the find_element_by_* methods. Replace driver.find_element_by_class_name('foo') with driver.find_element(By.CLASS_NAME, 'foo'). Don’t forget from selenium.webdriver.common.by import By. The same applies to find_element_by_id, find_element_by_xpath, find_element_by_css_selector, etc.

Should I catch AttributeError with try/except?

Generally no — AttributeError usually indicates a real bug in your code (wrong type, wrong API). Catching it hides the bug. The legitimate exception is duck typing patterns where you genuinely want to handle “this might or might not have method X”: try: x.close() except AttributeError: pass. But for normal code, fix the type mismatch instead of catching the error.

How often is this AttributeError reference updated?

New posts are added weekly as we encounter errors in real projects. Existing posts are revised every 6–12 months when major library versions ship (NumPy 2.x, pandas 2.x, Selenium 4 → 5, TensorFlow versions). This page was last refreshed in May 2026.

Related error categories

AttributeError is one of 5 hubs in our Python error reference cluster — 765+ documented fixes total. If your error isn’t about a missing attribute, jump to the right hub below:

• TypeError Reference — 220+ Python and JavaScript TypeError fixes (NoneType, not callable, JSON serializable, JS undefined)
• ModuleNotFoundError Reference — 198+ “No module named X” Python import errors
• ValueError Reference — 100+ library-specific Python ValueError fixes (pandas, NumPy, sklearn, TensorFlow)
• ImportError Reference — 67+ “cannot import name X from Y” version-conflict fixes
• Python Tutorial — beginner-to-intermediate Python lessons

About this AttributeError reference

This AttributeError reference has been built since 2015 by PIES Information Technology Solutions in Binalbagan, Negros Occidental, Philippines. Every post in the collection started as a real error encountered in production code — by us, by our clients, or by readers who emailed us with their tracebacks. Used by 12,000+ Python developers monthly across the Philippines, India, the United States, and beyond. If your AttributeError isn’t here, send the full traceback to our contact form and we’ll add it.