How to speed up data retrieval and assign a value?

Question

For example, I have two data frames

df1:

        0        1           2           3 
0      Name     Unit      Attribute     Date
1      a        A           xxy         xxx
2      b        B           xyx         xxx
3      c        C           xxx         xxx
4      d        D           yxx         xxx
5      e        E           yxy         yyy

df2:

        0        1        2       
      Name     Unit      Date
0      a        F        xxx
1      b        G        xxx
2      e        H        xxx
3      f        I        xxx

I want to overwrite entries in df1 with their corresponding entries in df2.

For example, replace df1.loc [5,3] with df2.loc [2,2]. That is, for rows with the same "name", rewrite df1 in the same column if it is in df2.

I'm currently doing this stupidly:

def find_column_num(key, df_name, start_row, stop_row, start_column, stop_column):
    for i in range(start_row,stop_row+1):
        for j in range(start_column, stop_column+1):
            if df_name.loc[i,j]== key:
                column_num_with_key = j
                return column_num_with_key
                break 

for i in range(0,len(df1.index)):
    for ii in range(0,len(df2.index)):
       if df1.loc[i,0] == df2.loc[ii,0]:
          for j in range(0,len(df1.columns)):
              if df1.loc[0,j] in df2.loc[0,:]:
                   df1.set_value(i,j, df2.loc[ii,find_column_num(df1.loc[0,j],df2,0,0,0,len(df2.columns))]

I am not proud of it. I did a little research and came up with replacing '=' with set_value (), which helps. I sincerely hope to hear other suggestions. The size of the actual problem is 200 rows and 30 columns. Thus, going through all the for loops takes 20 seconds.

Answer 1

IIUC, merge fillna. . .

 df1

  Name Unit Attribute Date
0    a    A       xxy  xxx
1    b    B       xyx  xxx
2    d    C       xxx  xxx
3    e    D       yxx  xxx
4    e    E       yxy  xxx

df2

  Name Unit Date
0    a    F  xxx
1    b    G  xxx
2    e    H  xxx
3    f    H  xxx

---

out = df1[['Name', 'Attribute']].merge(df2, how='left').fillna(df1)
out

  Name Attribute Unit Date
0    a       xxy    F  xxx
1    b       xyx    G  xxx
2    d       xxx    C  xxx
3    e       yxx    H  xxx
4    e       yxy    H  xxx