Split Data Into Multiple columns Based On rows of one Column

Here's an R answer. If you create a grouping vector using cumsum on the presence of "3C" (or some other identifier for your groups, perhaps the "|"-character) you can then split and use the first item (one time) to any remaining items via R's recycling convention for dataframe definition:

dat <- read.table(text=txt)  # copied your data into txt
dat <- cbind(dat, grp=cumsum( grepl("3C", dat$V1) ))
#grepl pattern could have been "assembly" if that were more general

   do.call(rbind,  lapply( split(dat, dat$grp), 
           function(x) data.frame(
                            group=x[1,1], # first gets recycled
                            item=x[-1,1]) )  ) # the rest
                    group       item
1   3C-assembly|contig_93 ptg000037l
2   3C-assembly|contig_94 ptg000039l
3   3C-assembly|contig_95 ptg000043l
4.1 3C-assembly|contig_96 ptg000196l
4.2 3C-assembly|contig_96 ptg000060l
5.1 3C-assembly|contig_97 ptg000083l
5.2 3C-assembly|contig_97 ptg000083l
6.1 3C-assembly|contig_98 ptg000117l
6.2 3C-assembly|contig_98 ptg000005l
7.1 3C-assembly|contig_99 ptg000123l
7.2 3C-assembly|contig_99 ptg000123l
7.3 3C-assembly|contig_99 ptg0001232
7.4 3C-assembly|contig_99 ptg0001233

In python:

#Assuming the data is in pandas dataframe. I just created it:

import pandas as pd
a=[
"3C-assembly|contig_93 ptg000037l",
"3C-assembly|contig_94 ptg000039l",
"3C-assembly|contig_95 ptg000043l",
"3C-assembly|contig_96 ptg000196l",
"3C-assembly|contig_96 ptg000060l",
"3C-assembly|contig_97 ptg000083l",
"3C-assembly|contig_97 ptg000083l",
"3C-assembly|contig_98 ptg000117l",
"3C-assembly|contig_98 ptg000005l",
"3C-assembly|contig_99 ptg000123l",
"3C-assembly|contig_99 ptg000123l",
"3C-assembly|contig_99 ptg0001232",
"3C-assembly|contig_99 ptg0001233"]

a=pd.DataFrame(a, columns=["data"])

#Define Function to SPlit and Extract
def ExtractContig(Name):
    #Split Based on Space
    splitgroup=Name.strip().split(' ')
    contigselect = splitgroup[0]
    ptgselect=splitgroup[1]
    
    # Split Based on Underscore to get first column
    contig = contigselect.strip().split('_')[-1]    
    
    #Split Based on "g" of the string ptgxxxxxx
    ptg = ptgselect.strip().split('g')[-1]    
    return   contig,ptg

#Function Call and Collect Title for Each rows
a['data'].apply(lambda Name: ExtractContig(Name))

You can store and perform further analysis. The output in this case is:

0     (93, 000037l)
1     (94, 000039l)
2     (95, 000043l)
3     (96, 000196l)
4     (96, 000060l)
5     (97, 000083l)
6     (97, 000083l)
7     (98, 000117l)
8     (98, 000005l)
9     (99, 000123l)
10    (99, 000123l)
11    (99, 0001232)
12    (99, 0001233)
Name: data, dtype: object