如何将 DataFrame 的列名从字符串转换为整数

Question

In the following code, I read a string into a DataFrame, but even though the headers of the input string are numbers, they are read in as strings '1', '2' . Is there a way to read them in as numbers, or convert them to numbers afterwards?

import pandas as pd
from StringIO import StringIO


string_input = " 1 2\n10 0.1 0.2\n20 0.1 0.2"
data = pd.read_table(StringIO(string_input), sep='\s+')
print data
print data.columns

      1    2
10  0.1  0.2
20  0.1  0.2

Index([u'1', u'2'], dtype='object') # the columns names are of type str!!

Answer 1

You can do this as a post-processing step using astype(int) :

In [86]:
string_input = " 1 2\n10 0.1 0.2\n20 0.1 0.2"
data = pd.read_table(io.StringIO(string_input), sep='\s+')
print (data)
print (data.columns.astype(int))
​
      1    2
10  0.1  0.2
20  0.1  0.2
Int64Index([1, 2], dtype='int64')

personally I would prefer string columns as it becomes less ambiguous when indexing IMO when reading and writing code, as in doing df['col_name'] becomes a habit and when you have a default int64 index then df.loc[some_int] is unambiguous