迹忆客 专注技术分享

我们将介绍如何使用 insert 和 reindex 以不同的方法更改 pandas.DataFrame 列的顺序，例如以所需的顺序分配列名。

以新顺序在对 pandas.DataFrame 列排序

最简单的方法是用 columns 的列表重新分配 DataFrame，或者只是按照我们想要的顺序分配列名：

# python 3.x
import pandas as pd

df = pd.DataFrame(
    {
        "a": ["1", "2", "3", "4"],
        "b": [16, 7, 6, 16],
        "c": [61, 57, 16, 36],
        "d": ["12", "22", "13", "44"],
        "e": ["Green", "Red", "Blue", "Yellow"],
        "f": [1, 11, 23, 66],
    }
)
print(df)
df = df[["e", "c", "b", "f", "d", "a"]]
print("Rearranging ..................")
print("..............................")
print(df)

输出：

   a   b   c   d       e   f
0  1  16  61  12   Green   1
1  2   7  57  22     Red  11
2  3   6  16  13    Blue  23
3  4  16  36  44  Yellow  66
Rearranging ..................
..............................
        e   c   b   f   d  a
0   Green  61  16   1  12  1
1     Red  57   7  11  22  2
2    Blue  16   6  23  13  3
3  Yellow  36  16  66  44  4

在 Pandas 中用 insert 添加新列和位置

如果我们要创建新列，则可以将其插入到我们想要的任何位置：

# python 3.x
import pandas as pd

df = pd.DataFrame(
    {
        "a": ["1", "2", "3", "4"],
        "b": [16, 7, 6, 16],
        "c": [61, 57, 16, 36],
        "d": ["12", "22", "13", "44"],
        "e": ["Green", "Red", "Blue", "Yellow"],
        "f": [1, 11, 23, 66],
    }
)
print(df)
print("Inserting ..................")
print("..............................")
df.insert(0, "newColMean", df.mean(1))
print(df)

输出：

   newColMean  a   b   c   d       e   f
0   26.000000  1  16  61  12   Green   1
1   25.000000  2   7  57  22     Red  11
2   15.000000  3   6  16  13    Blue  23
3   39.333333  4  16  36  44  Yellow  66

Pandas 中给定顺序的 reindex 列

reindex 可以说是重新排列列的最有效方法：

# python 3.x
import pandas as pd

df = pd.DataFrame(
    {
        "a": ["1", "2", "3", "4"],
        "b": [16, 7, 6, 16],
        "c": [61, 57, 16, 36],
        "d": ["12", "22", "13", "44"],
        "e": ["Green", "Red", "Blue", "Yellow"],
        "f": [1, 11, 23, 66],
    }
)
print(df)
print("Rearranging ..................")
print("..............................")
df = df.reindex(columns=["a", "f", "d", "b", "c", "e"])
print(df)

输出：

   a   b   c   d       e   f
0  1  16  61  12   Green   1
1  2   7  57  22     Red  11
2  3   6  16  13    Blue  23
3  4  16  36  44  Yellow  66
Rearranging ..................
..............................
   a   f   d   b   c       e
0  1   1  12  16  61   Green
1  2  11  22   7  57     Red
2  3  23  13   6  16    Blue
3  4  66  44  16  36  Yellow