python - 在这种情况下如何通过匹配公共(public)值来生成新的数据框？

Question

我有两个这样的数据框:

data_2019_dict = {'state': ['Ohio', 'Texas', 'Pennsylvania', 'Nevada', 'New York', 'Nevada', 'Ohio', 'Virginia', 'Louisiana', 'Florida', 'Nevada'],
 'industry': ['Agriculture', 'Agriculture', 'Agriculture', 'Agriculture', 'Medicine', 'Medicine', 'Medicine', 'Medicine', 'Manufacture', 'Manufacture', 'Manufacture'],
 'value': [3.6, 3.2, 2.9, 2.4, 3.1, 1.5, 1.4, 0.9, 4.4, 2.0, 1.9]}

data_2020_dict = {'state': ['Kansas', 'Texas', 'California', 'Idaho', 'Nevada', 'Ohio', 'Virginia', 'Louisiana', 'Texas', 'Nevada'],
 'industry': ['Agriculture', 'Agriculture', 'Agriculture', 'Medicine', 'Medicine', 'Finance', 'Finance', 'Manufacture', 'Manufacture', 'Manufacture'],
'value': [2.3, 1.8, 1.6, 7.2, 5.9, 4.1, 0.2, 5.1, 2.3, 2.2]}

data_2019 = pd.DataFrame(data_2019_dict)
data_2020 = pd.DataFrame(data_2020_dict)

每个数据框都显示一年中哪些州在这些行业中表现良好。我想要生成但陷入困境的是:对于每个州来说，哪些行业在这两年中表现良好？生成的数据框将如下所示:

# Manually generated for illustration 
data_both_dict = {'state': ['Ohio', 'Texas', 'Pennsylvania', 'Nevada', 'Nevada', 'New York', 'Virginia', 'Louisiana', 'Florida', 'Kansas', 'California', 'Idaho'],
             'common_industry': ['', 'Agriculture', '', 'Medicine', 'Manufacture', '', '', 'Manufacture', '', '', '', ''],
             'common_industry_count': [0, 1, 0, 2, 2, 0, 0, 1, 0, 0, 0, 0]
}
data_both = pd.DataFrame(data_both_dict)

Answer 1

第一DataFrame.merge对于两列的公共(public)行，重命名列并按 Series.value_counts 添加计数和 Series.map :

df = (data_2019.merge(data_2020, on=['state','industry'])
               .rename(columns={'industry':'common_industry'}))
df['common_industry_count'] = df['state'].map(df['state'].value_counts())
df = df[['state','common_industry','common_industry_count']]
print (df)
       state common_industry  common_industry_count
0      Texas     Agriculture                      1
1     Nevada        Medicine                      2
2  Louisiana     Manufacture                      1
3     Nevada     Manufacture                      2

然后通过 concat 获取所有状态删除重复项 Series.drop_duplicates和 Series.to_frame 的一列 DataFrame :

both = pd.concat([data_2019['state'], data_2020['state']]).drop_duplicates().to_frame()
print (both)
          state
0          Ohio
1         Texas
2  Pennsylvania
3        Nevada
4      New York
7      Virginia
8     Louisiana
9       Florida
0        Kansas
2    California
3         Idaho

最后一次合并left连接并用 Series.fillna 替换缺失值:

df = both.merge(df, how='left')
df['common_industry_count'] = df['common_industry_count'].fillna(0).astype(int)
df['common_industry'] = df['common_industry'].fillna('')
print (df)
           state common_industry  common_industry_count
0           Ohio                                      0
1          Texas     Agriculture                      1
2   Pennsylvania                                      0
3         Nevada        Medicine                      2
4         Nevada     Manufacture                      2
5       New York                                      0
6       Virginia                                      0
7      Louisiana     Manufacture                      1
8        Florida                                      0
9         Kansas                                      0
10    California                                      0
11         Idaho                                      0