Create data frames for each unique permutation of groups of a Dataframe-CodePudding

Suppose I have the following Dataframe in Python:

     input_df = pd.DataFrame({
    'Previous': ['1000', '1000', 'latex', 'latex'], 
    'Ignore':[None, None, ['free'], ['free']], 
    'New': ['100', '200', 'nylon', 'cloth']})

I would like to generate the following four dataframes:

    df1  = pd.DataFrame({
        'Previous': ['1000','latex'],
        'Ignore': [None, ['free']],
        'New': ['100','nylon']})
    df2  = pd.DataFrame({
        'Previous': ['1000','latex'],
        'Ignore': [None, ['free']],
        'New': ['100','cloth']})
    df3  = pd.DataFrame({
        'Previous': ['1000','latex'],
        'Ignore': [None, ['free']],
        'New': ['200','nylon']})
    df4  = pd.DataFrame({
        'Previous': ['1000','latex'],
        'Ignore': [None, ['free']],
        'New': ['200','cloth']})

How can I accomplish this?

Edit: I have arrived at the following solution by modifying @TheMaster 's answer:

out=[pd.DataFrame(j) for j in c([i[1] for i in input_df.iterrows()], len(input_df['Previous'].unique())) if len(pd.DataFrame(j)['Previous'].unique()) == len(input_df['Previous'].unique())]

This solution keeps only the output where the 'Previous' column has all unique entries.

CodePudding user response：

Try combinations:

from itertools import combinations as c
out=[pd.DataFrame(j) for j in c([i[1] for i in df.iterrows()],2)]

Out[2]:
[  Previous  New
 0     1000  100
 1     1000  200,
   Previous    New
 0     1000    100
 2    latex  nylon,
   Previous    New
 0     1000    100
 3    latex  cloth,
   Previous    New
 1     1000    200
 2    latex  nylon,
   Previous    New
 1     1000    200
 3    latex  cloth,
   Previous    New
 2    latex  nylon
 3    latex  cloth]