Blog
-
Pandas Series duplicate index labels: When .loc returns a Series
The index in a Python Pandas series feels like a dict keys. But the keys can repeat:
-
Pandas Series indexing: Why .loc beats [] for retrieval
To retrieve from a Python Pandas series, you can use [] to retrieve items. But don’t! In a series, [] uses the index. In a data frame, [] uses the column names. Confusing! Besides, .loc does everything [] does, but with more options and flexibility.
-
Pandas set_axis: Replacing a Series index without mutating it
You can set a Python Pandas series index with set_axis. This returns a new series with the new index applied: s.index # still Index([‘a’, ‘b’, ‘c’], dtype=’str’)
-
Pandas Series index: Change it by assigning to s.index
Want to change the index of a Python Pandas series? Just assign to it:
-
Pandas Series index: Setting labels instead of RangeIndex
By default, a Python Pandas series has a RangeIndex, like a string or list: Set an index like this:
-
isinstance vs type(): The safer way to check Python types
Is a Python value of a particular type? It’s tempting to say: Far better to say: Why?– Works with subclasses– Second argument can be a tuple of possibilities
-
Python inheritance: When to use it and when to prefer composition
When should you use inheritance in Python? – An existing class mostly does what you want– Inheriting reduces your work — you only define differences– It’s an IS-A relationship, not just HAS-A– Otherwise: prefer composition. Remember: Inheritance is a tool, not a default.
-
Python metaclasses: How they work and when to use a decorator instead
What is a Python metaclass? – A subclass of type– Classes can be instances of a metaclass, rather than type– The metaclass’s __init__ sets attributes and methods Realistically, you should probably use a decorator instead of a metaclass. (But they’re fun to play with!)
-
Python metaclasses: How type.__new__ and type.__init__ create classes
A Python class is an instance of type. So type.__new__ and type.__init__ create the class and set attributes. Consider MyType, a subclass of type, setting some attributes in its __init__. MyType is a *metaclass*, used like this:
-
object and type in Python: The circular class hierarchy
Every Python class inherits (directly or not) from object. object is a class. So type(object) is type, the class of all classes. But type is a class, and inherits from object! So yes: object is an instance of type. And type inherits from object. Also: type(type) is type!