How to Resolve "TypeError: reduction operation 'argmax' not allowed for this dtype" in Pandas

The TypeError: reduction operation 'argmax' not allowed for this dtype in Pandas arises when you attempt to use functions like idxmax(), argmax(), idxmin(), or argmin() on columns that don't contain numeric data.

This guide explains the common causes of this error and provides solutions using pandas.to_numeric() and DataFrame.astype() to ensure your data is in the correct format.

Understanding the Error

The idxmax(), argmax(), idxmin(), and argmin() methods are designed to find the index or position of the maximum or minimum value in a numeric column. If the DataFrame column's data type is object (often representing strings), these operations can not be performed, resulting in the TypeError.

import pandas as pd

df = pd.DataFrame({
    'experience': ['5', '14', '7', '10'],
    'salary': [175.1, 180.2, 190.3, 205.4],
}, index=['Alice', 'Bob', 'Carl', 'Dan'])

print(df.dtypes)
print(df.idxmax())  # TypeError: reduction operation 'argmax' not allowed for this dtype

Output:

experience     object
salary        float64
dtype: object
Traceback (most recent call last):
  File "main.py", line 9, in <module>
    print(df.idxmax())
  File "/usr/lib/python3.8/site-packages/pandas/core/frame.py", line 8861, in idxmax
    indices = nanops.nanargmax(self.values, axis=axis, skipna=skipna)
  File "/usr/lib/python3.8/site-packages/pandas/core/nanops.py", line 66, in _f
    raise TypeError(
TypeError: reduction operation 'argmax' not allowed for this dtype

• The error occurs because the first column contains object datatypes (strings) while the salary column contains floats.
• Since strings are not directly comparable this will throw an exception.

Solutions

To solve this, convert the non-numeric columns to a numeric type before calling the desired method.

Converting Columns to Numeric with pandas.to_numeric()

The pandas.to_numeric() method is the most flexible way to convert a column to a numeric type:

import pandas as pd

df = pd.DataFrame({
    'experience': ['5', '14', '7', '10'],
    'salary': [175.1, 180.2, 190.3, 205.4],
}, index=['Alice', 'Bob', 'Carl', 'Dan'])

df['experience'] = pd.to_numeric(df['experience']) # Convert 'experience' to numeric

print(df.dtypes)
print(df.idxmax()) # Now it will execute

Output:

experience      int64
salary        float64
dtype: object
experience    Bob
salary        Dan
dtype: object

• The example converts all strings to int64, and then proceeds to call idxmax function, which works as expected now.

Converting All Object Columns to Numeric

To convert all object columns in the DataFrame you can use a for loop.

First use eq() with object to select the columns.

import pandas as pd

df = pd.DataFrame({
    'experience': ['5', '14', '7', '10'],
    'another': ['1', '3', '14', '5'],
    'salary': [175.1, 180.2, 190.3, 205.4],
}, index=['Alice', 'Bob', 'Carl', 'Dan'])

object_columns = df.columns[df.dtypes.eq(object)]

print(object_columns)
# Output: Index(['experience', 'another'], dtype='object')

df[object_columns] = df[object_columns].apply(
    pd.to_numeric,
    errors='coerce',
    axis=0
)

print(df.idxmax())

Output:

Index(['experience', 'another'], dtype='object')
experience     Bob
another       Carl
salary         Dan
dtype: object

• The dtypes attribute gets the type of every column in the dataframe.
• The DataFrame.apply is called to map all of the strings from all columns to numbers, throwing errors='coerce' to avoid the ValueError if a specific column contains strings that are not numeric.

Using DataFrame.astype()

The DataFrame.astype() method is another way to convert the columns to numeric types.

import pandas as pd

df = pd.DataFrame({
    'experience': ['5', '14', '7', '10'],
    'salary': [175.1, 180.2, 190.3, 205.4],
}, index=['Alice', 'Bob', 'Carl', 'Dan'])

df['experience'] = df['experience'].astype(int) # Explicitly cast to integer
print(df.idxmax())

Output:

experience    Bob
salary        Dan
dtype: object

• The call to DataFrame.astype(int) converts the columns into an int type.

Checking Data Types

Before calling any numerical operations such as idxmax, always check the datatypes of your DataFrame columns using df.dtypes:

import pandas as pd

df = pd.DataFrame({
    'experience': ['5', '14', '7', '10'],
    'salary': [175.1, 180.2, 190.3, 205.4],
}, index=['Alice', 'Bob', 'Carl', 'Dan'])
print(df.dtypes) # Displays the data type of each column

Output:

experience     object
salary        float64
dtype: object

This helps identify columns that need conversion.

Conclusion

The TypeError: reduction operation 'argmax' not allowed for this dtype in Pandas is easily resolved by ensuring your data is in the correct numeric format before performing the method calls.

• The most reliable way to do this is to use pandas.to_numeric() to convert values and handle any potential bad data.
• Use df.dtypes to check that your columns have been properly converted.