How do I update some values of my target column in a dataframe by matching its multiple columns with one column in another dataframe in python?

Question

I have two pandas dataframes:

• df1
• df2

df1 looks like this:

a     f1     f2     f3     target
1     aa     bb     cc     NaN
2     xx     yy     zz     NaN
more rows....

df2 looks like this:

b     f1     target
3     bb     450
4     xx     220
more rows....

The desired output for df1 is this:

a     f1     f2     f3     target
1     aa     bb     cc     450
2     xx     yy     zz     220
more rows....

This is my question:

How do I update df1["target"] with values from df2["target"] by matching df2["f1"] with df1["f1"] or df1["f2"] or df1["f3"] .

The merge and update functions don't work for me because I need to match multiple columns. Other solutions I saw were also for matching ALL multiple columns, which is not the case for my data. (I only need to match one column with ANY multiple columns.)

Thanks.

Answer 1

one approach would be to create a dictionary and apply it to first column if possible, second if possible etc.

d = dict(zip(df2.f1.values,df2.target.values))
df1.target = df1.apply(lambda r: d.get(r['f1'],d.get(r['f2'],d.get(r['f3']))), axis=1)

Answer 2

Consider iteratively merging into an appended dataframe that merges to original one:

# PAIRWISE MERGE FIELDS
merge_cols = [[i, 'f1'] for i in list(df1.columns[1:len(df1.columns)-1])]
# LIST OF INNER JOIN MERGES
dfs = [pd.merge(df1, df2, left_on=m[0], right_on=m[1], how='inner') for m in merge_cols]

# APPEND DFS
stackdf = pd.concat(dfs)[['a','target_y']]
finaldf = stackdf.merge(df1, on='a')\
                 .assign(target=stackdf['target_y'].values)\
                 .drop(['target_y'], axis=1)

print(finaldf)
#    a  f1  f2  f3  target
# 0  2  xx  yy  zz     220
# 1  1  aa  bb  cc     450