CODEWITHSUNDEEP
pythonpandasstringdataframe
Guilherme Reis
Guilherme Reis

Reputation: 205

How to get a substring based on position in pandas?

I have a pandas dataframe as shown below:

df =pd.DataFrame({'String':['JAIJAOD','ERJTD','AJDIDO','AJDIDO'],'Position':[5,2,nan,4]})

I am trying to get a third column that shows what is the letter of first column that represents the number in column Position. The dataframe should be something like

df = pd.DataFrame({'String':['JAIJAOD','ERJTD','AJDIDO','AJDIDO'],'Position':[5,2,nan,4],'Letter':['O','J',nan,'D']})

I have tried the following code, however, the output is not exactly what I want since the final table has some mistakes regarding the third column.

third = []
for i, n in zip(df['String'],df['Position']):
      if n >0: #I thought it because the column Position have just floats
         third.append(i[int(n)]
      else:
         third.append(np.nan)
df['Third'] = pd.Series(third)

Upvotes: 1

Views: 1127

Answers (2)

David Erickson
David Erickson

Reputation: 16683

You can apply a lambda x: function to the required input columns simultaneously, by applying to the dataframe, passing x to the dataframe and axis=1. For each row, my approach takes a slice of each value in String depending on the corresponding value in the Position column:

df =pd.DataFrame({'String':['JAIJAOD','ERJTD','AJDIDO','AJDIDO'],'Position':[5,2,np.nan,4]})
df['Letter'] = df[df['Position'].notnull()].apply(lambda x: x['String'][int(x['Position'])],axis=1)
df
Out[1]: 
    String  Position Letter
0  JAIJAOD       5.0      O
1    ERJTD       2.0      J
2   AJDIDO       NaN    NaN
3   AJDIDO       4.0      D

Upvotes: 0

BENY
BENY

Reputation: 323326

Let us try

df['Letter'] = [x[int(y)] if y==y else np.nan  for x , y in zip(df.String,df.Position) ]
['O', 'J', nan, 'D']

Upvotes: 2

Related Questions