Advanced Pandas Tutorial: Multi‑Index and apply Method

Goal : Deeply understand advanced Pandas features.

Learning Content : Multi‑level indexing and the apply method.

Code Examples :

import pandas as pd
# Create example dataset
data = {
    '姓名': ['张三', '李四', '王五', '张三', '赵六', '李四'],
    '部门': ['销售部', '市场部', '技术部', '销售部', '财务部', '市场部'],
    '销售额': [120, 150, 130, 160, 140, 170],
    '成本': [80, 90, 100, 110, 120, 130]
}
df = pd.DataFrame(data)
print(f"示例数据集: \n{df}")
# Set multi‑level index
df.set_index(['部门', '姓名'], inplace=True)
print(f"设置多级索引后的数据集: \n{df}")
# Select data by multi‑level index
selected_data = df.loc[('销售部', '张三')]
print(f"按多级索引选择的数据: \n{selected_data}")
# Partial index selection
selected_data_partial = df.loc['销售部']
print(f"按部分索引选择的数据: \n{selected_data_partial}")
# Reset index
df_reset = df.reset_index()
print(f"重置索引后的数据集: \n{df_reset}")
# Sort index
df_sorted = df.sort_index(level=['部门', '姓名'])
print(f"排序多级索引后的数据集: \n{df_sorted}")
# Apply method examples
def add_ten(x):
    return x + 10
df['销售额加十'] = df['销售额'].apply(add_ten)
print(f"对 '销售额' 列应用函数后的数据集: \n{df}")

def process_row(row):
    row['利润'] = row['销售额'] - row['成本']
    return row
df_processed = df.apply(process_row, axis=1)
print(f"对整个 DataFrame 应用函数后的数据集: \n{df_processed}")
# Lambda expression
df['销售额平方'] = df['销售额'].apply(lambda x: x ** 2)
print(f"使用 lambda 表达式后的数据集: \n{df}")
# Multiple aggregations
df_aggregated = df['销售额'].agg(['sum', 'mean', 'max', 'min'])
print(f"对 '销售额' 列应用多个函数后的结果: \n{df_aggregated}")

Practice : Use the multi‑level index to perform complex data operations, following the same code steps as above.

Summary : By completing this exercise you should have mastered Pandas' multi‑level indexing and the apply method, enabling flexible data management, querying, and custom transformations; further advanced Python data‑processing techniques will be explored in upcoming sessions.