Dividing a single column of dataframe into multiple columns based on char length

[darpInd]

Hi Guys,
I am importing a dataset from a url using below function:

## Import from the url
import pandas as pd
dftemp=pd.read_csv("http://users.stat.ufl.edu/~winner/data/agedeath.dat")
print (dftemp)
print (type(dftemp))

#####
It works fine and give me a result like:

Output:
        aris km    21      1 
0     aris  km  21      2 
1     aris  km  21      3 
2     aris  km  21      4 
 6183  sovr va  100   1439 
6184  sovr va  101   1440  
[6185 rows x 1 columns] <class 'pandas.core.frame.DataFrame'> 
#########

my objective:
I want to divide this single column in multiple columns -- eg.. for first row ----first column is for "aris km", 2nd for 21 and third for 1 . This should be based on the characters, like first value "aris km or sovr va is confined within 6 characters -- then 21 is occupying between 8-14 characters and last one is occupying 16 to 18 characters.

If I use split function, I will be able to break it using the space and thus 'aris" and 'km" will also be different columns while it should be same column.

How to break in this way?

[jefsummers]

You posted a while ago and no responses, so I am going to stab at it. First, you need to add "header = None" to your import statement, as the first row is being read as the column names, which it isn't. Next, for your question, I would use the "one hot encoding" technique (can read about it multiple sites) using .map . Others will hopefully be able to give you better ideas, but this can give you a start.

[scidam]

There is Series.str.extract method in Pandas which can
split data into columns. In your case it could be applied as follows:

dftemp.iloc[:,0].str.extract(r'([a-zA-Z\s]+)(\d+)(\d+)')