如何定义返回DataFrame索引的函数，实现冒号运算符切片

Got it, let's build that reusable function you need to skip writing manual loc slices every time. Here's a clean, flexible implementation that matches exactly what you're looking for:

import pandas as pd

def get_c_values(df, a_index, b_index=slice(None)):
    """
    Slice a multi-index DataFrame (indexed by ['A', 'B']) to fetch values from column 'C'
    
    Args:
        df: Your target pandas DataFrame
        a_index: Value for the first level index ('A')
        b_index: Value for the second level index ('B'), defaults to slice(None) (equivalent to ':')
    
    Returns:
        The corresponding value(s) from column 'C'
    """
    return df.loc[a_index, b_index]['C']

# Example usage with your DataFrame
df = pd.DataFrame({'A':['a', 'a', 'b'], 'B':[1,2,3], 'C':[1.1, 2.2 ,3.3]})
df = df.set_index(['A', 'B'])

# Get single value (matches df.loc['a',1]['C'])
print(get_c_values(df, 'a', 1))  # Output: 1.1

# Get multiple values (matches df.loc['a',:]['C'])
print(get_c_values(df, 'a'))  # Output:
# B
# 1    1.1
# 2    2.2
# Name: C, dtype: float64

How it works: #

• We set b_index to default to slice(None)—this is pandas' internal way of representing the : slice operator. It lets you omit the second argument when you want all values tied to a specific 'A' index.
• The function mirrors your manual workflow: it uses df.loc[a_index, b_index] to slice the multi-index, then pulls the 'C' column directly.

Optional: Add Robust Error Handling #

If you want the function to handle cases where the requested index doesn't exist gracefully, wrap the return logic in a try-except block:

def get_c_values(df, a_index, b_index=slice(None)):
    try:
        return df.loc[a_index, b_index]['C']
    except KeyError:
        return f"No matching values found for A='{a_index}', B='{b_index}'"

This replaces a raw KeyError with a friendly message when the index combination you're asking for doesn't exist in the DataFrame.

内容的提问来源于stack exchange，提问作者Niklas