How to put elements of a pandas index as values of a dictionary?

Question

I’m struggling with an error in pandas which is driving me crazy.

I want to build a dictionary which extracts some data from a pandas df:

Index_col  col1 
P1         F1-R1
P2         F1-R1
P3         F1-R1
P4         F1-R1
P5         F1-R2
P6         F1-R2
P7         F1-R2
P8         F1-R2
(etc)

Would give:

{'F1-R1': ['P1', 'P2', 'P3', 'P4'],
 'F1-R2': ['P5', 'P6', 'P7', 'P8']}

However the following code:

dic = dict.fromkeys(df.col1.unique(), [])
for idx, row in df.iterrows():
    dic[row["col1"]].append(idx)

Produces

{'F1-R1': ['P1', 'P2', 'P3', 'P4', 'P5', 'P6', 'P7', 'P8']],
 'F1-R2': ['P1', 'P2', 'P3', 'P4', 'P5', 'P6', 'P7', 'P8']}

I can’t figure ut why :/. Does someone has an answer (or another way to proceed) ?

Asked By: Micawber

||

Source

Answer 1

Do this:

import pandas as pd

data = {'Index_col': ['P1', 'P2', 'P3', 'P4', 'P5', 'P6', 'P7', 'P8'],
        'col1': ['F1-R1', 'F1-R1', 'F1-R1', 'F1-R1', 'F1-R2', 'F1-R2', 'F1-R2', 'F1-R2']}
df = pd.DataFrame(data)

dic = {key: [] for key in df.col1.unique()}
for idx, row in df.iterrows():
    dic[row["col1"]].append(row["Index_col"])
print(dic)

which gives

{'F1-R1': ['P1', 'P2', 'P3', 'P4'], 'F1-R2': ['P5', 'P6', 'P7', 'P8']}

Answered By: Serge de Gosson de Varennes

Answer 2

Just group by col1 and collect lists with converting the result into dict:

d = df.groupby('col1')['Index_col'].apply(list).to_dict()

{'F1-R1': ['P1', 'P2', 'P3', 'P4'], 'F1-R2': ['P5', 'P6', 'P7', 'P8']}

Answered By: RomanPerekhrest

Answer 3

Another possible solution. which uses dictionary comprehension:

{x: df.loc[df['col1'].eq(x),'Index_col'].to_list() for x in df['col1'].unique()}

Output:

{'F1-R1': ['P1', 'P2', 'P3', 'P4'], 'F1-R2': ['P5', 'P6', 'P7', 'P8']}

Answered By: PaulS

How to put elements of a pandas index as values of a dictionary?

Question:

Answers: