是否可以将numpy字符串数组转换为就地熊猫数据框？

Question

As you can see, the array has been copied.

import pandas as pd
import numpy as np
foo = np.array([['a', 'b'],['c', 'd']],dtype='U32')
df_foo = pd.DataFrame(data=foo, copy=False)
foo[0,0]='different'
print(df_foo.values)

Returns

[['a' 'b']
 ['c' 'd']]

Furthermore, foo has dtype='<U32' while df_foo.values has dtype=object . I don't need a solution to the problem, I'm more curious as to whether it's possible and if not, why not?

Answer 1

Yes, just remove copy=False and then it won't be a copy. So just replace

df_foo = pd.DataFrame(data=foo, copy=False)

with

df_foo = pd.DataFrame(data=foo)

It worked for me.