将pandas整数系列转换为字符串的最有效方法？

Question

Is .astype(str) the most efficient way to convert a series of ints to a series of strings? It seems rather slow so I wanted to ask.

Answer 1

I tried a few things and found a faster way using numpy:

setup = """
import pandas, numpy
s = pandas.Series(numpy.random.randint(1,10,(100)))
"""

>>> timeit.timeit('s.astype(str)', setup=setup, number=10000)
3.33058500289917
>>> timeit.timeit('s.apply(str)', setup=setup, number=10000)
3.572000026702881
>>> timeit.timeit('s.apply(lambda x: str(x))', setup=setup, number=10000)
3.821247100830078
>>> timeit.timeit('s.values.astype(numpy.str)', setup=setup, number=10000)
0.08432412147521973

As you can see, accesing the numpy array with values and using the astype call is over 40x faster than the next fastest method.