How to access the values of a pandas DataFrame with multiple headers by column name when iterating o-CodePudding

With a basic DataFrame with one header, one can iterate over the rows and access the values by column names:

import pandas as pd

df = pd.DataFrame(columns=['header1_column1', 'header1_column2'])

df['header1_column1'] = range(2)
df['header1_column2'] = range(2)

print(df)

   header1_column1  header1_column2
0                0                0
1                1                1

for index, row in df.iterrows():
    print(row['header1_column1'])

0
1

However, with a DataFrame that has multiple headers, iterating over the rows and accessing the values by column names yields an output with some overhead:

df = pd.DataFrame(columns=[['header1_column1', 'header1_column2'],
                           ['header2_column1', 'header2_column2']])

df['header1_column1'] = range(2)
df['header1_column2'] = range(2)

print(df)

  header1_column1 header1_column2
  header2_column1 header2_column2
0               0               0
1               1               1

for index, row in df.iterrows():
    print(row['header1_column1'])

header2_column1    0
Name: 0, dtype: int64
header2_column1    1
Name: 1, dtype: int64

How can I eliminate the overhead and have the same output as in the first case?

CodePudding user response：

I think you need select by tuple for MultiIndex columns:

for index, row in df.iterrows():
    print(row[('header1_column1','header2_column1')])
0
1