如何在Pandas中传入内置函数作为参数？通用空值处理函数报错求解

Got it, let's work through this issue step by step. The AttributeError: 'Series' object has no attribute 'action' happens because you're trying to call a method named action directly on the Series—but action is just a string parameter (like "mean"), not an actual built-in method of the Pandas Series object.

Here are two clean, scalable ways to properly pass and use these Pandas methods as parameters:

Method 1: Use getattr() to Dynamically Access Methods #

The getattr() function lets you fetch an object's method using a string, which is exactly what we need here. We'll also handle the "delete" case separately since it's a DataFrame-level operation, not a fill operation.

import pandas as pd

def fill_null(df, col_name, action):
    # Make a copy to avoid modifying the original DataFrame
    df_tmp = df.copy()
    target_col = df_tmp[col_name]

    if action in ["mean", "median"]:
        # Dynamically get the method (e.g., target_col.mean()) and call it
        fill_value = getattr(target_col, action)()
        df_tmp[col_name] = target_col.fillna(fill_value)
    elif action == "drop":
        # Remove rows with null values in the target column
        df_tmp = df_tmp.dropna(subset=[col_name])
    else:
        raise ValueError("Invalid action. Choose from 'mean', 'median', 'drop'")
    
    return df_tmp

if __name__ == '__main__':
    df_sal = pd.read_excel("sample_sal.xlsx")
    # Test mean filling
    df_mean_filled = fill_null(df_sal, "Salary", "mean")
    print("Mean-filled Salary Data:\n", df_mean_filled)

    # Test null row deletion
    df_dropped_nulls = fill_null(df_sal, "Salary", "drop")
    print("\nData after removing null Salary rows:\n", df_dropped_nulls)

Method 2: Use a Dictionary to Map Actions to Logic #

This approach makes your code more readable and easier to extend (e.g., adding a "mode" fill later). We map each action string to a lambda function that executes the corresponding logic.

import pandas as pd

def fill_null(df, col_name, action):
    df_tmp = df.copy()
    target_col = df_tmp[col_name]

    # Map action strings to their respective operations
    action_logic = {
        "mean": lambda: target_col.fillna(target_col.mean()),
        "median": lambda: target_col.fillna(target_col.median()),
        "drop": lambda: df_tmp.dropna(subset=[col_name])
    }

    if action not in action_logic:
        raise ValueError("Invalid action. Choose from 'mean', 'median', 'drop'")
    
    result = action_logic[action]()
    # Handle fill vs drop operations differently
    if action in ["mean", "median"]:
        df_tmp[col_name] = result
        return df_tmp
    else:
        return result

if __name__ == '__main__':
    df_sal = pd.read_excel("sample_sal.xlsx")
    df_median_filled = fill_null(df_sal, "Salary", "median")
    print("Median-filled Salary Data:\n", df_median_filled)

Key Explanation #

Your original code tried to call df[col_name].action()—this treats action as a literal method name, not the string value you passed to the function. Using getattr() or a dictionary mapping lets you translate the string parameter into the actual Pandas method you want to run.

内容的提问来源于stack exchange，提问作者Sidhartha