使用 pandas to_csv 时删除 header 中的前导逗号

Question

默认情况下 to_csv 会像

一样写入 CSV

,a,b,c
0,0.0,0.0,0.0
1,0.0,0.0,0.0
2,0.0,0.0,0.0

但我希望它这样写：

a,b,c
0,0.0,0.0,0.0
1,0.0,0.0,0.0
2,0.0,0.0,0.0

如何实现？我无法设置 index=False 因为我想保留索引。我只想删除前导逗号。

df = pd.DataFrame(np.zeros((3,3)), columns = ['a','b','c'])
df.to_csv("test.csv") # this results in the first example above.

Answer 1

只需为您的索引设置一个名称：df.index.name = 'blah'。此姓名将作为名字出现在 headers.

中

import numpy as np
import pandas as pd

df = pd.DataFrame(np.zeros((3,3)), columns = ['a','b','c'])
df.index.name = 'my_index'
print(df.to_csv())

产量

my_index,a,b,c
0,0.0,0.0,0.0
1,0.0,0.0,0.0
2,0.0,0.0,0.0

但是，如果（根据您的评论）您希望 headers 中有 3 个 coma-separated 名称，而 csv 行中有 4 个 coma-separated 值，则您必须手工制作。但是它不符合任何 csv 标准格式。

Answer 2

或者，尝试重置索引，使其成为数据框中的一列，名为 index。这也适用于多个索引。

df = df.reset_index()
df.to_csv('output.csv', index = False)

Answer 3

可以先只写没有索引的列，然后在追加模式下写没有 header 的数据：

df = pd.DataFrame(np.zeros((3,3)), columns = ['a','b','c'], index=list('XYZ'))

pd.DataFrame(columns=df.columns).to_csv("test.csv", index=False)
#alternative for empty df
#df.iloc[:0].to_csv("test.csv", index=False)
df.to_csv("test.csv", header=None, mode='a')

df = pd.read_csv("test.csv")
print (df)
     a    b    c
X  0.0  0.0  0.0
Y  0.0  0.0  0.0
Z  0.0  0.0  0.0

使用 pandas to_csv 时删除 header 中的前导逗号

Remove leading comma in header when using pandas to_csv

python

dataframe

export-to-csv

pandas