如何合并两个数据框，但一个 dataframe 有多个符合条件的单元格？

Question

Assuming we have two dataframes:

    df1 = pd.DataFrame(
        [["foo",10], ["bar",20]], 
        columns=["1", "2"], 
        index=["x", "y"]
    )
    df2 = pd.DataFrame(
        [["foo",10,20], ["bar",20,30], ["foo",10,30]],
        columns=["1", "2", "3"],
        index=["x", "y", "z"]
    )

This would give us this:

output of df1 & df2

If I were to merge these data with a condition on two columns:

df3 = pd.merge(df1, df2, how='left', on=['1','2'])

This would give us this:

output of merged df3

If I wanted the mean of the values that match the condition from df2 to be output into df3, how would I go about doing that? (So instead of having two rows of foo & 10, I would only have one row of foo & 10, with the values for the third column being an average of the two rows that match the condition). I provided a picture below for clarity:

wanted output of merged df3

Answer 1

If you want mean then try:

df2 = df2.groupby(['1','2'], as_index=False).mean()
df3 = pd.merge(df1, df2, how='left', on=['1','2'])