Convert pandas column values based on groupings of values

I have a pandas columns with values 1.0, 2.0, 3.0, 4.0, and 5.0 like below:

0       5.0
1       2.0
2       3.0
3       3.0
4       5.0
       ... 
1039    5.0
1040    1.0
1041    2.0
1042    4.0
1043    1.0

I want rows with values 1.0 or 2.0 to all have a value of 1.0, 3.0 and 4.0 to become 2.0, and 5.0 to become 3.0. How could I re-assign the values based on these groupings. I was thinking np.where() at first but now I’m not sure how to implement that with np.where() logic because that seems like it would be better suited for conversion to a binary variable. Maybe just masking with .loc()?

Thanks.

>Solution :

Given your pattern, use simple arithmetics: add 1, get the floor division by 2:

df['new'] = df['col'].add(1).floordiv(2)

Or use cut:

df['new'] = pd.cut(df['col'], [0, 2, 4, 6], labels=[1, 2, 3])

Note that cut will give you a categorical type.

Example:

      col  new
0     5.0  3.0
1     2.0  1.0
2     3.0  2.0
3     3.0  2.0
4     5.0  3.0
1039  5.0  3.0
1040  1.0  1.0
1041  2.0  1.0
1042  4.0  2.0
1043  1.0  1.0