Create an if-elif scenario statement as a dictionary-CodePudding

I have the following statement, if col1="string1" and col2="string3" then newcol="newstring" and if col1="stringX" and col2="stringY" then newcol="newstringZ", how do I frame this into a dictionary so that I can apply map function to use the dictionary to check and insert new column for a dataframe?

df = pd.DataFrame(
{
        'col1': ['string1', 'string1', 'string3', 'stringX', 
                '','stringX'],
        'col2': ["stringY","string3","stringY",np.nan,"stringY","stringY"]
}
)

After applying map function, the end result should look like this

CodePudding user response：

Let's use 'merge' with a dataframe constructor:

dict_map = {'col1':['string1','stringX'], 'col2':['string3', 'stringY'], 'newcol':['newstring', 'newstringZ']}

df_map = pd.DataFrame(dict_map)

df.merge(df_map, how='left')

Output:

      col1     col2      newcol
0  string1  stringY         NaN
1  string1  string3   newstring
2  string3  stringY         NaN
3  stringX      NaN         NaN
4           stringY         NaN
5  stringX  stringY  newstringZ

CodePudding user response：

You can set the dict with tuples of (col1, col2) as keys and values for newcol as values. Then, turn col1, col2 as a tuple (composite key) by .apply() on axis=1. Then use .map() function for the mapping, as follows:

mapper = {('string1', 'string3') : 'newstring', 
          ('stringX', 'stringY') : 'newstringZ'}

df['newcol'] = df[['col1', 'col2']].apply(tuple, axis=1).map(mapper)

Result:

print(df)


      col1     col2      newcol
0  string1  stringY         NaN
1  string1  string3   newstring
2  string3  stringY         NaN
3  stringX      NaN         NaN
4           stringY         NaN
5  stringX  stringY  newstringZ