I have a large data frame that I want to merge with another dataset. In order to do so I need to get the names of individuals into a certain format. The function below converts the name in 'column' to the desired format (mostly) and stores it in newColumn.

My question is, is there a better (faster and/or more pythonic) way to do this?

The main aim is to transform full names into surname and initials, such as:

  • Novak Djokovic = Djokovic N.
  • Jo-Wilfred Tsonga = Tsonga J.W.
  • Victor Estrella Burgos = Estrella Burgos V.
  • Juan Martin Del Potro = Del Potro J.M.

def convertNames(df,column, newColumn):
    df[newColumn] = 'none'

    for player in df[column]:

        names = player.split(' ')

        if len(names) == 2:

            if (len(names[0].split('-')) > 1):
                newName = names[1]+' '+names[0].split('-')[0][0]+'.'+names[0].split('-')[1][0]+'.'
                newName = names[1]+' '+names[0][0]+'.'

        elif len(names) == 3:
            newName = names[1]+' '+names[2]+' '+names[0][0]+'.'

            newName = names[2]+' '+names[3]+' '+names[0][0]+'.'+names[1][0]+'.'

        df[newColumn][df[column] == player] = newName

    return df
  • 4
    \$\begingroup\$ Your code doesn’t match your sample output. The code converts “Victor Estrella Burgos” into “Estrella Burgos V.”, not “Burgos V.E.” Which represents the desired behaviour: the sample output or the code? \$\endgroup\$
    – AJNeufeld
    Dec 13, 2018 at 4:10
  • 1
    \$\begingroup\$ What about Alex De Minaur? \$\endgroup\$
    – stefan
    Dec 13, 2018 at 8:34
  • \$\begingroup\$ @AJNeufeld, Sorry, that was me just trying to copy and past everything. You are right. The sample code produces the sample code produces the correct output. I have corrected the question. \$\endgroup\$ Dec 14, 2018 at 4:33
  • \$\begingroup\$ @Stefan, the desired result for Alex De Minaur is De Minaur A. \$\endgroup\$ Dec 14, 2018 at 4:34
  • 1
    \$\begingroup\$ What about John Patrick Smith? Any rule converting Alex De Minaur to De Minaur A. will convert John Patrick Smith to Patrick Smith J.. Spaniards mostly have two surnames, others mostly one. Any simple approach will fail. \$\endgroup\$
    – stefan
    Dec 14, 2018 at 18:48

1 Answer 1


You are doing way too much split()-ing. You split on ‘-‘, and if you find the length of the split is greater than 1, you split on the ‘-‘ twice more, to get the first and the second part of the hyphenated name. Split once, and save the result in a list, and access the list elements!

You are doing too much in convertNames(). It would be better to create a convertName() method, which just processes the player name into the desired form. Then you could call that method from convertNames().

def convertName(player):
    names = player.split(' ')

    if len(names) == 2:
        names[0:1] = names[0].split('-', 1)

    surname = min(len(names)-1, 2)

    return ' '.join(names[surname:]) + ' ' + ''.join(name[0]+'.' for name in names[:surname])

# Test data
for player in ('Novak Djorkovic', 'Jo-Wilferd Tsonga', 'Victor Estrella Burgos', 'Juan Martin Del Potro'):
    print(player, ':', convertName(player))

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.