import pyperf
runner = pyperf.Runner()

setup="""
from dataclasses import dataclass, asdict
@dataclass
class D:
	a : dict
	b: list
	
d=D({'1': 1}, [1,2,3])
asdict(d)
"""

runner.timeit(name=f"micro benchmark", stmt=f"hasattr(o, '__str__')", setup='import os; o=type(os)')
runner.timeit(name=f"micro benchmark: misssing attr", stmt=f"hasattr(o, '_no_such_attr')", setup='import os; o=type(os)')
runner.timeit(name=f"asdict on dataclass", stmt=f"asdict(d)", setup=setup)

micro benchmark: Mean +- std dev: [main] 42.1 ns +- 1.1 ns -> [pr] 44.1 ns +- 1.8 ns: 1.05x slower
micro benchmark: misssing attr: Mean +- std dev: [main] 256 ns +- 6 ns -> [pr] 37.6 ns +- 0.7 ns: 6.80x faster
asdict on dataclass: Mean +- std dev: [main] 8.87 us +- 0.17 us -> [pr] 6.61 us +- 0.05 us: 1.34x faster

Geometric mean: 2.06x faster