Pandas 多层索引操作的实现

df = pd.DataFrame(
    [[621,582,621,591],
     [615,585,616,597],
     [581,574,579,585],
     [596,565,584,561]])
# 数组
# 每个数组对应着一个层级的索引值
arrays = [['北京','北京','上海','上海'],['北大','清华','上交','复旦']]
mindex = pd.MultiIndex.from_arrays(arrays, names=['城市','大学'])
# 给df行索引赋值
df.index = mindex

# dataframe
# 创建一个dataframe，方式与元组类似，每个元组对应一对多级索引值
frame = pd.DataFrame([('北京','北大'),('北京','清华'),('上海','上交'),('上海','复旦')])
mindex = pd.MultiIndex.from_frame(frame, names=['城市','大学'])
# 给df行索引赋值
df.index = mindex

df = pd.DataFrame(
    [[621,582,621,591],
     [615,585,616,597],
     [581,574,579,585],
     [596,565,584,561]])
# 元组
# 每个元组是对应着一对多级索引
tuples = [('北京','北大'),('北京','清华'),('上海','上交'),('上海','复旦')]
mindex = pd.MultiIndex.from_tuples(tuples, names=['城市','大学']) 
# 给df行索引赋值
df.index = mindex

# product笛卡尔积

city = ['北京', '上海']
college = ['北大','清华','上交','复旦']
mindex1 = pd.MultiIndex.from_product([city,college], names=['城市','大学'])
mindex1

MultiIndex(levels=[['上海', '北京'], ['上交', '北大', '复旦', '清华']],
           codes=[[1, 1, 1, 1, 0, 0, 0, 0], [1, 3, 0, 2, 1, 3, 0, 2]],
           names=['城市', '大学'])

# product生成column列索引
year = ['2022','2023']
pro = ['计算机','土木']
mcol = pd.MultiIndex.from_product([year,pro], names=['年份','专业'])
 
# 对df的行索引、列索引赋值
df.index = mindex
df.columns = mcol
display(df)

def index_info():
    address =os.path.join('.','index_def.xlsx')
    criterion = pd.read_excel(address)
    multi_idx1=criterion['分类'].unique()
    multi_idx2=criterion['二级分类'].unique()
    
    idx1 = [multi_idx1[0]]*3+[multi_idx1[2]]*3+[multi_idx1[3]]*16
    idx2 = [multi_idx2[0]]*3+['-']*3+[multi_idx2[2]]*6+[multi_idx2[3]]*6+[multi_idx2[4]]*4

    criterion.index=pd.MultiIndex.from_arrays([idx1,idx2,criterion.index+1],names=['一级分类','二级分类','编号'])
    ctr_df = criterion.iloc[:,np.r_[3,5]].fillna('-')
    return ctr_df
ctr_df = index_info()
ctr_df