concat之后，值更改为float，而不是int（尝试在concat之前将其转换。）

Question

我首先尝试将信息（系列）更改回int，然后在我将其更改回cncat之后，它又更改回float>

import pandas as pd

df1= pd.DataFrame({'a': [1,2,3, None,5,6,7,8,9,10],
                  'code': ['a{}'.format(i+1) for i in range(10)]})
df2 = pd.DataFrame({'code': ['a{}'.format(i+1) for i in range(5)]})

df1['a'] = pd.to_numeric(df1['a'], errors='coerce')
df1_temp = df1.dropna(subset=['a'])
df1_temp_2 = df1[df1['a'].isnull()]
df1_temp['a'] = df1_temp['a'].astype(int)
df1_a = pd.concat([df1_temp, df1_temp_2], axis=0)
df1_a.sort_index(axis=0)
df1_a

输出：

    a   code
0   1.0 a1
1   2.0 a2
2   3.0 a3
3   NaN a4
4   5.0 a5
5   6.0 a6
6   7.0 a7
7   8.0 a8
8   9.0 a9
9   10.0    a10

预期：

    a   code
0   1   a1
1   2   a2
2   3   a3
3   NaN a4
4   5   a5
5   6   a6
6   7   a7
7   8   a8
8   9   a9
9   10  a10

有什么好建议吗？越短越好

Answer 1

在熊猫0.24+和nullable integer data type中有可能：

df1['a'] = pd.to_numeric(df1['a'], errors='coerce').astype('Int64')
print (df1)
     a code
0    1   a1
1    2   a2
2    3   a3
3  NaN   a4
4    5   a5
5    6   a6
6    7   a7
7    8   a8
8    9   a9
9   10  a10

编辑：如果要转换^{}中的整数（如果已将整数写入文件）：

df1= pd.DataFrame({'a': [1,2,3, None,5,6,7,8,9,10],
                  'code': ['a{}'.format(i+1) for i in range(10)]})
df1['a'] = pd.to_numeric(df1['a'], errors='coerce').astype('Int64')

df1.to_csv('test.csv', index=False)

df = pd.read_csv('test.csv', dtype={'a':'Int64'})
print (df)
     a code
0    1   a1
1    2   a2
2    3   a3
3  NaN   a4
4    5   a5
5    6   a6
6    7   a7
7    8   a8
8    9   a9
9   10  a10