Btukfyl

I have object type columns that have categorical values example 15-16 Years, 17-23 Years ..... I have converted them to category and then cat.codes. However,the coding values start from 0 for the first group 0-4 years and I want the coding to start from 1 ie 0-4->1, 5-12-> and @@->NaN
The suggested solution of using dictionary mapping still has issues. Following is mcve

import pandas as pd

data = ['0-4 Years', '5-12 Years','13-18 Years', '19-21 Years','22-25 Years','26-29 Years','30-35 Years',

        '36-41 Years','42-45 Years','46-49 Years','50-55 Years', '56-63 Years']

df = pd.DataFrame(data,columns=['Age'],dtype=object)

df['Age']=df['Age'].astype('category')

cats = dict(enumerate(df['Age'].cat.categories, 2))

df['Age']=df['Age'].cat.codes.map(cats).astype('category')

df['Age']

and here is the output, as you can see if I change the enumeration start other than 0, there is an issue with values as NaN. Secondly the column is not coded either:

df['Age']

0             NaN

1     36-41 Years

2             NaN

3             NaN

4       0-4 Years

5     13-18 Years

6     19-21 Years

7     22-25 Years

8     26-29 Years

9     30-35 Years

10    42-45 Years

11    46-49 Years

Name: Age, dtype: category

Categories (9, object): [0-4 Years, 13-18 Years, 19-21 Years, 22-25 Years, ..., 30-35 Years, 36-41 Years, 42-45 Years, 46-49 Years]

How to fix this.

You can create you own dictionary that maps codes and categories with:

cats = dict(enumerate(df['Age'].cat.categories, 1))

And use this dictionary to map it in the dataframe

df['Age'].cat.codes.map(cats).astype('category')