如何在Pandas DataFrame中用Americas行减SouthAmerica行并新增NorthAmerica行?
用Pandas的groupby和diff实现区域数据差值计算并添加新行
原始数据
创建DataFrame的代码:
import pandas as pd df = pd.DataFrame({ "country": ["Americas", "Americas","Europe", "Europe", "SouthAmerica", "SouthAmerica"], "year": [2000, 2005, 2000, 2005, 2000, 2005], "planted": [900, 800, 300, 100, 500, 300], "regrowth": [300, 400, 500, 200, 200, 100] })
实际数据展示:
country year planted regrowth 0 Americas 2000 900 300 1 Americas 2005 800 400 2 Europe 2000 300 500 3 Europe 2005 100 200 4 SouthAmerica 2000 500 200 5 SouthAmerica 2005 300 100
需求
计算Americas与SouthAmerica在planted和regrowth列的差值(Americas数值 - SouthAmerica数值),把结果作为NorthAmerica行添加到原DataFrame末尾,最终输出如下:
country year planted regrowth 0 Americas 2000 900 300 1 Americas 2005 800 400 2 Europe 2000 300 500 3 Europe 2005 100 200 4 SouthAmerica 2000 500 200 5 SouthAmerica 2005 300 100 6 NorthAmerica 2000 400 100 7 NorthAmerica 2005 500 300
解决方案(基于groupby+diff)
步骤1:筛选目标区域并排序
先把Americas和SouthAmerica的数据筛出来,按年份升序、国家名称降序排序——这样每个年份组里,Americas会排在SouthAmerica前面,方便后续计算差值。
target_df = df[df['country'].isin(['Americas', 'SouthAmerica'])].sort_values(['year', 'country'], ascending=[True, False])
步骤2:分组计算差值
按year分组后,对planted和regrowth列用diff(-1)方法。diff(-1)是计算当前行和下一行的差值,刚好能得到每个年份里Americas减SouthAmerica的结果。
diff_result = target_df.groupby('year')[['planted', 'regrowth']].diff(-1)
步骤3:构造NorthAmerica数据
把差值结果里的空值(NaN)删掉,添加country列(值为NorthAmerica)和year列,最后调整列顺序和原DataFrame一致:
north_america = diff_result.dropna().reset_index(drop=True) north_america['country'] = 'NorthAmerica' north_america['year'] = [2000, 2005] north_america = north_america[['country', 'year', 'planted', 'regrowth']]
步骤4:合并到原DataFrame
用concat把新构造的NorthAmerica数据追加到原DataFrame后面,重置索引得到连续的序号:
final_df = pd.concat([df, north_america], ignore_index=True)
完整可运行代码
import pandas as pd # 原始数据 df = pd.DataFrame({ "country": ["Americas", "Americas","Europe", "Europe", "SouthAmerica", "SouthAmerica"], "year": [2000, 2005, 2000, 2005, 2000, 2005], "planted": [900, 800, 300, 100, 500, 300], "regrowth": [300, 400, 500, 200, 200, 100] }) # 筛选并排序目标区域数据 target_df = df[df['country'].isin(['Americas', 'SouthAmerica'])].sort_values(['year', 'country'], ascending=[True, False]) # 按年份分组计算差值 diff_result = target_df.groupby('year')[['planted', 'regrowth']].diff(-1) # 构造NorthAmerica数据 north_america = diff_result.dropna().reset_index(drop=True) north_america['country'] = 'NorthAmerica' north_america['year'] = [2000, 2005] north_america = north_america[['country', 'year', 'planted', 'regrowth']] # 合并得到最终结果 final_df = pd.concat([df, north_america], ignore_index=True) print(final_df)
关键逻辑说明
- 排序是为了保证每个年份内,要被减数(Americas)排在减数(SouthAmerica)前面,这样
diff(-1)就能直接得到我们需要的差值。 groupby('year')确保差值计算是在同一年份内进行,不会出现跨年份的错误计算。diff(-1)和默认的diff()(当前行减上一行)方向相反,这里用它刚好匹配我们“前减后”的需求。
内容的提问来源于stack exchange,提问作者krawall
相关产品推荐
相关产品推荐

