Change Specific Word in Pandas-CodePudding

I have a dataframe

city = pd.DataFrame({'id': [1,2,3,4],
                        'city': ['NRTH CAROLINA','NEW WST AMSTERDAM','EAST TOKYO','LONDON STH']})

How can I change NRTH to NORTH, WST to WEST, and STH to SOUTH, so the output will be like this

id   city
1    NORTH CAROLINA
2    NEW WEST AMSTERDAM
3    EAST TOKYO
4    LONDON STH

CodePudding user response：

Hello, Arthur! I have defined mapping_dict where you can define any other words that you want to change.

For changing them I made a separate function for mapping city names.

import pandas as pd
city = pd.DataFrame({'id': [1,2,3,4],
                        'city': ['NRTH CAROLINA','NEW WST AMSTERDAM','EAST TOKYO','LONDON STH']})
mapping_dict = {'NRTH':'NORTH','WST':'WEST','STH':'SOUTH'}
def mapping_words(city_name):
    updated_name = ""
    for word in city_name.split():
        if word in mapping:
            updated_name  = mapping[word] " "
        else:
            updated_name  = word " "
    return updated_name.strip()

city['city'] = city['city'].apply(lambda x: mapping_words(x))

I hope this may help you. Thanks!

CodePudding user response：

Let's define a replace dictionary first then use Series.replace(regex=True) to replace by the word boundary of the dictionary key.

import re

d = {
    'NRTH': 'NORTH',
    'WST': 'WEST',
    'STH': 'SOUTH'
}


df['city'] = df['city'].replace({rf"\b{re.escape(k)}\b":v for k, v in d.items()}, regex=True)

print(df)

   id                city
0   1      NORTH CAROLINA
1   2  NEW WEST AMSTERDAM
2   3          EAST TOKYO
3   4        LONDON SOUTH