Combining pandas with different columns

Question

I have two pandas dataframes that look like this:

df1
    id                  10  20          30          
0   2020-04-01-10001    0.0 0.000000    0.000000    
1   2020-04-01-10003    0.0 0.026587    0.053174    
2   2020-04-01-10005    0.0 0.030884    0.061768    
3   2020-04-01-1001     0.0 0.035875    0.071751    
4   2020-04-01-1003     0.0 0.041673    0.083346    
...

df2
                    10      20      30
id                                  
2020-04-01-10001    0.15    0.17    0.18    
2020-04-01-10003    0.55    0.57    0.61    
2020-04-01-10005    0.36    0.37    0.38    
2020-04-01-1001     0.00    0.00    0.02    
2020-04-01-1003     0.00    0.02    0.04
...

I want to use rows of df1 to update the rows of df2, so I am trying to use df2.update(df1). Unfortunately, this doesn't change the rows of df2. I suspect that it has something to do with the extra numbers on the left side of df1. I'm not sure what those are. I also noticed that the two tables give different results when I run df.to_dict:

'10': {0: 0.0, 1: 0.0, 2: 0.0, 3: 0.0, 4: 0.0},,...

{'10': {'2020-04-01-10001': 0.15,
  '2020-04-01-10003': 0.55,
  '2020-04-01-10005': 0.36,
  '2020-04-01-1001': 0.0,
  '2020-04-01-1003': 0.0},...

How would I go about converting df1 to the format of df2?

Answer 1

Use set_index with replace and fillna :

import numpy as np

df1 = df1.set_index('id').replace(0, np.NaN).fillna(df2)

                    10        20        30
id                                        
2020-04-01-10001  0.15  0.170000  0.180000
2020-04-01-10003  0.55  0.026587  0.053174
2020-04-01-10005  0.36  0.030884  0.061768
2020-04-01-1001   0.00  0.035875  0.071751
2020-04-01-1003   0.00  0.041673  0.083346