pandas.api.typing.Expression#

class pandas.api.typing.Expression(func, repr_str, needs_parenthese=False)[source]#

Class representing a deferred expression evaluated against a DataFrame.

Expressions are initially created via pd.col and can be composed using arithmetic, comparison, and any method or attribute that the underlying object supports. This includes NumPy ufuncs, indexing, and the boolean operators &, |, and ~.

Expressions are solely intended for use in expression-based APIs, such as DataFrame.assign and DataFrame.loc, where they are evaluated against the DataFrame they are passed to.

This is not meant to be instantiated directly. Instead, use pandas.col().

See also

pandas.col: Create a deferred column reference.

Notes

The attributes that are available depend on the expression and where it is used. Attribute access is itself deferred: it is applied to the result of evaluating the expression, so an attribute that is not supported by that result raises when the expression is evaluated.

Examples

>>> df = pd.DataFrame({"a": [1, 2], "b": [3, 4]})
>>> df.assign(c=pd.col("a") + pd.col("b"))
   a  b  c
0  1  3  4
1  2  4  6