Hello, I am trying to merge two data sets via:
df1=pd.dataframe({'company name':['a','b','c'], 'analyst 1 name':['tom','mike',np.nan], 'analyst 2 name':[np.nan,'alice',np.nan], 'analyst 3 name':['jane','steve','alex']}) df2=pd.dataframe({'company name':['a','b','c'], 'score 1':[3,5,np.nan], 'score 2':[np.nan,1,np.nan], 'score 3':[6,np.nan,11]}) df_desire=pd.dataframe({'company name':['a','a','b','b','b','c'], 'analyst':['tom','jane','mike','alice','steve','alex'], 'score':[3,6,5,1,np.nan,11]})
Basically, df1 contains the analyst name and df2 contains the score assigned by the analyst. I'm trying to merge the two into df_desire. The method of reading the two tables is: for company a, it is covered by two people, namely tom and jane, who assign 3 and 6 respectively. Note that although steve covers company b, I intentionally specify the score as na for robustness purposes.
What I did was:
pd.concat([df1.melt(id_vars='company name',value_vars=['analyst 1 name','analyst 2 name','analyst 3 name']),\ df2.melt(id_vars='company name',value_vars=['score 1','score 2','score 3'])],axis=1)
I'm looking for a more elegant solution.
Attempt:
x = ( df1.set_index("company name") .stack(dropna=false) .reset_index(name="name") .drop(columns="company name") ) y = df2.set_index("company name").stack(dropna=false).reset_index(name="score") print( pd.concat([x, y], axis=1)[["company name", "name", "score"]] .dropna(subset=["name", "score"], how="all") .reset_index(drop=true) )
Print:
company name name score 0 A Tom 3.0 1 A Jane 6.0 2 B Mike 5.0 3 B Alice 1.0 4 B Steve NaN 5 C Alex 11.0
The above is the detailed content of Merge from width to length. For more information, please follow other related articles on the PHP Chinese website!