KeyError:

Python KeyError has a particularly confusing variant with JSON data. json.loads() always produces string keys in the resulting dictionary, even when the JSON object had integer-like keys written by some serializers. If code previously did json.dumps({1: 'value'}), the resulting JSON string {"1": "value"} deserializes back with string key '1', not integer 1. Accessing data[1] raises KeyError because data['1'] is the actual key. HTTP APIs that return JSON with numeric IDs as object keys — a pattern in some REST APIs — consistently produce this confusion. The response {"123": {...}} is a valid JSON object with string key '123', requiring data['123'] or data.get('123'). A separate edge case: dictionaries created from namedtuple._asdict() or dataclasses.asdict() produce string keys, but database result rows from SQLite's sqlite3.Row or SQLAlchemy's Row behave as case-insensitive mappings in some configurations. Converting to a plain dict(row) normalizes to lowercase keys, causing KeyError when code later accesses data['UserName'] (now only data['username'] works). Always use dict.get(key, default) for keys whose presence you cannot guarantee, and if key in d before access in critical paths.