From dataframe to a nested dictionary

Question

I am experiencing some hard time trying in creating a nested dictionary from a dataframe in python and would really appreciate your help.This is what the data looks like.

	tag	id	cat
0	PA14_00030	GO:0005524	F
1	PA14_00030	GO:0006281	P
2	PA14_00050	GO:0003677	F
3	PA14_00050	GO:0003918	F
4	PA14_00050	GO:0005524	F
5	PA14_00050	GO:0006265	P
6	PA14_00060	GO:0016746	F
7	PA14_00070	GO:0005975	C
8	PA14_00080	GO:0009055	C

I want to create a nested dictionary that will look something like this:

{'C':{'PA14_00080':{'GO:0009055'},'PA14_00070':{'GO:0005975'}},
'F':{'PA14_0003': {'GO:0005524'}, 'PA14_00050': {'GO:0003677', 'GO:0003918','GO:0005524'}}, 
'P':{PA14_00050:{GO:0006265}, PA14_00030:{GO:0006281}}}

Thank you for your help.

Answer 1

If I understand you correctly, you can do:

out = {
    k: {
        row["tag"]: set(g.loc[g["tag"] == row["tag"], "id"].to_list())
        for _, row in g.iterrows()
    }
    for k, g in df.groupby("cat")
}

print(out)

Prints:

{
    "C": {"PA14_00070": {"GO:0005975"}, "PA14_00080": {"GO:0009055"}},
    "F": {
        "PA14_00030": {"GO:0005524"},
        "PA14_00050": {"GO:0003918", "GO:0005524", "GO:0003677"},
        "PA14_00060": {"GO:0016746"},
    },
    "P": {"PA14_00030": {"GO:0006281"}, "PA14_00050": {"GO:0006265"}},
}