Argthtjtr

My dataframe contains these columns

ID   Address1   Address1-State   Address1-City  Address2  Address2-State  Address2-City   Address    State      City

 1    6th street   MN              Mpls

 2                                               15th St      MI           Flint

 3                 MA              Boston          Essex St   NY            New York

 4    7 street SE  MN              Mpls           8th St      IL             Chicago

Now I want to populate Address field in such a way that if Address1 is blank,populate Address2 and the state city fields of Address2

In the above case final dataframe would be like this

 ID     Address   State    City

  1      6th street   MN    Mpls

  2      15th St      MI    Flint

  3      Essex St     NY     New York

  4     7 street SE  MN     Mpls

Currently,I am doing this

def fill_add(address1,address2):

if address1!='':

    address=address1



elif address1=='' and address2!='':

    address=address2

elif address1=='' and address2=='':

     address=''



return address





def fill_add_apply(df):

df['Address']=df.apply(lambda row:fill_add(row['Address1'],row['Address2']),axis=1)

Do I have to do the same for all the other columns?Is there a better way?

Just to clarify, in ID=3 Address,State,City should be "Essex St NY New York" because Address 1 is blank so it should choose Address2 and the city and state of Address2.
In short, if Address1 is blank it should choose Address2,Address2-State and Address2-City,even if Address1-State and Address1-City are not blank.

1st modify your column , then using groupby +first

df=df.replace('',np.nan)#prepare for first 



df.columns=df.columns.str.replace('d+','')

df.columns=df.columns.str.split('-').str[-1]

newdf=df.groupby(level=0,axis=1).first()

newdf.loc[df.iloc[:,1].isnull(),:]=df.groupby(level=0,axis=1).last()

newdf

Out[40]: 

       Address      City  ID State

0  6th street       Mpls   1    MN

1      15th St     Flint   2    MI

2    Essexb St  New York   3    NY

3  7 street SE      Mpls   4    MN

import numpy as np



df=df.replace('',np.nan)



addr_1=['ID','Address1','Address1-State','Address1-City']

addr_2=['ID','Address2','Address2-State','Address2-City']



new_df=pd.DataFrame(df[addr_1].values.copy(),columns=['ID','Address','State','City'])



new_df.loc[new_df['Address'].isnull(),:]=df.loc[df['Address1'].isnull(),addr_2].values



#print(new_df)

    ID  Address     State   City

0   1   6th street  MN      Mpls

1   2   15th St     MI      Flint

2   3   Essex St    NY      New York

3   4   7 street SE MN      Mpls

(Given you do not have any duplicate index)

Select indices you want to fill with Adress1:

Address1_index = df.loc[!df.Address1.empty() and !df.Address1-State.empty() and !df.Address1-City.empty()].index

then put Address1 data in your desired columns:

df.loc[Adress1_index, ["Adress", "State", "City"]] = df.loc[Adress1_index, ["Adress1", "Adress1-State", "Adress1-City"]]

Now select indices you want to fill with address2:

Address2_index = df.loc[df.Adress1.empty() or df.Adress1-State.empty() or df.Adress1-City.empty()].index

then fill these also:

df.loc[Adress2_index, ["Adress", "State", "City"]] = df.loc[Adress2_index, ["Adress2", "Adress2-State", "Adress2-City"]]

Remove columns that you do not want:

df.drop(["Address1", "Adress1-State", "Adress1-City", "Address2", "Adress2-State", "Adress2-City"], axis = 1, inplace = True)