Merge three columns into one taking into account priority preference

I have the following DataFrame in pandas:

I want to merge the columns town, city and village into one. In case more than 1 column contains a value other than NaN, the resulting column will be assigned the value in this order of preference: village > town > city. Example:

>Solution :

Let us define the order of columns then select those columns and use backfill along columns axis

order = ['village', 'town', 'city']
df['merged'] = df[order].bfill(axis=1).iloc[:, 0]

   code       town          village              city           merged
0     1    Brunete              NaN               NaN          Brunete
1     2        NaN  Cabrera d'Anoia               NaN  Cabrera d'Anoia
2     3        NaN              NaN         Barcelona        Barcelona
3     4  Zarzalejo              NaN            Madrid        Zarzalejo
4     7    Melilla              NaN           Melilla          Melilla
5     8  Cartagena           Galifa  Region of Murcia           Galifa