当数据框架没有标题时，其 Dataframe.columns 属性采用数字索引。否则，它使用字符串。所以，只要检查第一列标签的类型即可。

import pandas as pd
import io
def has_header(df):
    return isinstance(df.columns[0], str)
csv=u"""col1,col2,col3
5,2,7
4,9,6
7,3,1"""
df1 = pd.read_csv(io.StringIO(csv))
print(df1.head())
if has_header(df1):
    print("Dataframe 1 has header")
else:
    print("Dataframe 1 doesn't have header")
csv=u"""5,2,7
4,9,6
7,3,1"""
df2 = pd.read_csv(io.StringIO(csv), header=None)
print(df2.head())
if has_header(df2):
    print("Dataframe 2 has header")
else:
    print("Dataframe 2 doesn't have header")
df3= pd.read_csv(io.StringIO(csv))
print(df3.head())
if has_header(df3):
    print("Dataframe 3 has header")
else:
    print("Dataframe 3 doesn't have header")
df4 = pd.read_csv(io.StringIO(csv), header='infer')
print(df4.head())
if has_header(df4):
    print("Dataframe 4 has header")
else:
    print("Dataframe 4 doesn't have header")
以下是上述代码产生的输出。
   col1  col2  col3
0     5     2     7
1     4     9     6
2     7     3     1
Dataframe 1 has header
   0  1  2
0  5  2  7
1  4  9  6
2  7  3  1
Dataframe 2 doesn't have header
   5  2  7
0  4  9  6
1  7  3  1
Dataframe 3 has header
   5  2  7