Updating dataframe column with values from a different dataframe

Question

I have two dataframes and I want to update one of the dataframe based on matching id's from the second one, but only on 1 column. Both dataframes contain other columns that I am not concerned about and should not be updated.

For example:

df1

id    name  ...
123   city a
456   city b
789   city c
789   city c
456   city b
123   city a
.
.
.

so on and so forth

df2

id    name  ...
123   City A
456   City B
789   City C
.
.
.

So the resulting df_new should be:

id    name  ...
123   City A
456   City B
789   City C
789   City C
456   City B
123   City A
.
.
.

Right now what I'm doing is:

df2 = df2.set_index('id')
df1 = df.set_index('id')
df1.update(df2)
df1.reset_index(inplace=True)
df2.reset_index(inplace=True)

So I tried to write this into a function like this:

def replace_names(df1, df2, 'id'):
    df2 = df2.set_index('id')
    df1 = df.set_index('id')
    df1.update(df2)
    df1.reset_index(inplace=True)
    df2.reset_index(inplace=True)
    return df

However, this seems to return a dataframe of NoneType when used as a function.

I would also like to change this to a method where instead of reindexing two dataframes and resetting them, if there is a way to pick an id and names column on two dataframes as variables and just apply a function like this only on those two columns, without updating or changing anything else in the rest of the columns on either dataframes.

Any help would be appreciated!

Answer 1

Change your function to:

def replace_names(df1, df2, idxCol = 'id', srcCol = 'name', dstCol = 'name'):
    df1 = df1.set_index(idxCol)
    df1[dstCol].update(df2.set_index(idxCol)[srcCol])
    return df1.reset_index()

Parameters:

• df1 - the DataFrame with column to be updated.
• df2 - the DataFrame with the source column.
• idxCol - name of the "join" column.
• srcCol - name of the source column (in df2 ).
• dstCol - name of the destination column (in df1 ).

Then call it:

df_new = replace_names(df1, df2)

Your code failed because:

• In df1 = df.set_index('id') you refer to df (an external variable, despite having both DataFrames specified as parameters).
• In df1.update(df2) you update all columns, whereas you want to update only one column.
• Your function returned df - a variable without any connection to what this function did so far (maybe it was an empty DataFrame).

Answer 2

   final= df1.merge(df2, on=['id'],how='inner')

then you can drop column that which is not required