一个选项是使用combine_first：

cols = ["run_id", "time"]
​
out = (
        df1.set_index(cols)
              .combine_first(df2.set_index(cols))
              .reset_index().sort_values(by=cols)
           [["time", "run_id", "weight", "totalizer"]]
       )

输出：

print(out)

    time run_id  weight  totalizer
0    0.0     H1   500.0        NaN
1    0.5     H1     NaN      100.0
2   10.0     H1     NaN      200.0
3   24.0     H1   400.0        NaN
4   40.0     H1     NaN      300.0
5   48.0     H1   300.0        NaN
6   60.0     H1     NaN      400.0
7    0.0     H2   900.0        NaN
8    0.5     H2     NaN      900.0
9    5.0     H2     NaN     1000.0
10  24.0     H2   800.0        NaN
11  35.0     H2     NaN     1100.0
12  48.0     H2   700.0        NaN
13  70.0     H2     NaN     1200.0

您可以简单地在您已经完成的操作之后添加行：

mergedf = df1.merge(df2, how='outer') # your current code
mergedf.sort_values(['run_id', 'time']) # add this

在此阅读更多信息：https://stackoverflow.com/a/17141755/2650341

你可以用Pandas的merge_ordered

df_merged=pd.merge_ordered(df1,df2, on=['run_id','time'])