From a previous question I asked, I wanted to create and sort a certain Parent/Child hierarchy recursively.
With the huge help of one member, he find me a solution to complete this work.

But, yeah there is a but, The input file is a .csv file, with 18 columns.

The previous solutions only uses two columns (Parent and Child) and the output file just get these two columns + the column we’ve created to represent the hierarchy.

Here is the link of the previous topic:
/questions/78364917/sort-hierarchic-parent-child-list-in-python

My goal was to simply add my new column (created with the previous answer) and keep all the original DATA.

The problem is : The data from the original .csv file doesn’t have any unique ID, so I can’t merge old and new Dataframe together.

I’ve also tried to add the columns manually, here is the code :

def make_hierarchy(g):
    G = nx.from_pandas_edgelist(g, create_using=nx.DiGraph,
                                source='REF_ARTICLE_PERE', target='REF_ARTICLE_FILS', **edge_attr=True**)
    
    def dfs_with_level(node, level, order, parent_order, max_depth):
        # Here I tried to modify the hierarchy_data squeletton, but no success
        hierarchy_data = [(g.name, node, level, parent_order + f'{order:0{max_depth}d}')]
        children = list(G.successors(node))
        if children:
            for i, child in enumerate(children, start=1):
                hierarchy_data.extend(dfs_with_level(child, level + 1, i, parent_order + f'{order:0{max_depth}d}' + '.', max_depth))
        return hierarchy_data
    
    hierarchy_data = []
    for node in G.nodes:
        if not list(G.predecessors(node)):
            max_depth = len(str(len(G)))
            hierarchy_data.extend(dfs_with_level(node, 1, 1, '', max_depth))
    # Here I tried to add columns in the dataframe, but they told me 4 cols passed, needed 5 somthing like this
    df_hierarchy = pd.DataFrame(hierarchy_data, columns=['REF_PRODUIT', 'REF_ARTICLE_FILS', 'Level', 'sorted_order'])

    # Add underscores to represent hierarchy levels visually (reverse order)
    max_level = df_hierarchy['Level'].max()
    df_hierarchy['Level'] = df_hierarchy['Level'].apply(lambda x: '_' * (x - 1) + str(x))
    
    return df_hierarchy


df = pd.read_csv('input.csv')

out = df.groupby('REF_PRODUIT', group_keys=False).apply(make_hierarchy)

out.to_csv('output.csv', index=False)

I set edge_attr=Yes to keep all the cols in the graph, but the problem is coming after this step.

While checking all the nodes, I would like to keep the data of the actual row, and add it to the df_hierarchy.

In the debugger I see the data in :

G.adj.values But it’s all the data of all rows, not of only the current row.
And even if I find a way to get the data of the current row (to add it to the new DF) I’m not sure if I’m able to add it …

I’m new to all of this Data manipulation, I’m a student, every help would be appreciated.