1、層次索引

1.1 定義

在某一個方向擁有多個（兩個及兩個以上）索引級別，就叫做層次索引。

通過層次化索引，pandas能夠以較低維度形式處理高緯度的數據

通過層次化索引，可以按照層次統計數據

層次索引包括Series層次索引和DataFrame層次索引

1.2 Series的層次索引

import numpy as npimport pandas as pds1 = pd.Series(data=[99, 80, 76, 80, 99],    index=[['2017', '2017', '2018', '2018', '2018'], ['張伊曼', '張巧玲', '張詩詩', '張思思', '張可可']])print(s1)

1.3 DataFrame的層次索引

# DataFrame的層次索引df1 = pd.DataFrame({ 'year': [2016, 2016, 2017, 2017, 2018], 'fruit': ['apple', 'banana', 'apple', 'banana', 'apple'], 'production': [10, 30, 20, 70, 100], 'profits': [40, 30, 60, 80,10],})print("df1===================================")print(df1)df2 = df1.set_index(['year', 'fruit'])print("df2===================================")print(df2)print("df2.index===================================")print(df2.index)print("df2.sum(level='year')===================================")print(df2.sum(level='year'))print("df2.mean(level='fruit')===================================")print(df2.mean(level='fruit'))print("df2.sum(level=['year', 'fruit'])===================================")print(df2.sum(level=['year', 'fruit']))

2、取值的新方法

ix是比較老的方法 新方式是使用iloc loc

iloc 對下標值進行操作 Series與DataFrame都可以操作

loc 對索引值進行操作 Series與DataFrame都可以操作

2.1 Series

# # 取值的新方法s1 = pd.Series(data=[99, 80, 76, 80, 99],    index=[['2017', '2017', '2018', '2018', '2018'], ['張伊曼', '張巧玲', '張詩詩', '張思思', '張可可']])print("s1=================================")print(s1)print("s1.iloc[2]=================================")print(s1.iloc[2])print("s1.loc['2018']['張思思']=================================")print(s1.loc['2018']['張思思'])

2.2 DataFrame

df1 = pd.DataFrame({ 'year': [2016, 2016, 2017, 2017, 2018], 'fruit': ['apple', 'banana', 'apple', 'banana', 'apple'], 'production': [10, 30, 20, 70, 100], 'profits': [40, 30, 60, 80,10],})print("df1===================================")print(df1)print("舊方法獲取值===================================")print("df1['year'][0]===================================")print(df1['year'][0])print("df1.ix[0]['year']===================================")print(df1.ix[0]['year'])print("新方法獲取值===================================")print("df1.iloc[0][3]===================================")print(df1.iloc[0][3])print("df1.loc[0]['year']===================================")print(df1.loc[0]['year'])