How to create a dictionary of series with an index from a dataframe in python

Question

I have the following dataframe

df = pd.DataFrame({'col_1': ['Animal','Deer','Sheep','Fish','Vehicle','Truck','Car','Mineral','Gold','Silver'], 
                   'col_2': ['Animal',0.5,0.25,0.25,'Vehicle',0.5,0.5,'Mineral',0.75,0.25],
                   'Group': [0,0,0,0,1,1,1,2,2,2]})

I want to create a dictionary of series. I want column 1 to be the index, column 2 the value and Group to specify the series. EG the name (key) for the first series would be "Animal" and it should look like:

I've tried the following. But it's not right, I'm getting a dictionary of dataframes instead of series and the headers are in the first row.

dict_of_series = {i: df.loc[df.group == i, ['col_1', 'col_2']] for i in range(1, df.group.iat[-1])}

Can you put your expected output

Tbaki
– Tbaki

2021-12-09 13:25:22 +00:00
Commented Dec 9, 2021 at 13:25 — Tbaki
– Tbaki, Commented Dec 9, 2021 at 13:25

jezrael · Accepted Answer · 2021-12-10 09:41:43Z

1

Use dictionary comprhension for loop by groupby object with DataFrame.set_axis for set columnsnames by first row per groups, remove first row and last column by indexing in DataFrame.iloc and last remove columns names in DataFrame.rename_axis :

dict_of_series = {g['col_2'].iat[0]: 
                  g.set_axis(g.iloc[0], axis=1).iloc[1:, :-1].rename_axis(None, axis=1) 
                  for i, g in df.groupby('Group')} 


print (dict_of_series['Animal'])
  Animal Animal
1   Deer    0.5
2  Sheep   0.25
3   Fish   0.25

print (dict_of_series['Vehicle'])
  Vehicle Vehicle
5   Truck     0.5
6     Car     0.5

print (dict_of_series['Mineral'])
  Mineral Mineral
8    Gold    0.75
9  Silver    0.25

For Series use DataFrame.set_index before solution and also change iloc for select last column to Series and last Series.rename_axis:

df = df.set_index('col_1')

dict_of_series = {g['col_2'].iat[0]: 
                  g.set_axis(g.iloc[0], axis=1).iloc[1:, 0].rename_axis(None)
                  for i, g in df.groupby('Group')} 


print (dict_of_series['Animal'])
Deer      0.5
Sheep    0.25
Fish     0.25
Name: Animal, dtype: object

edited Dec 10, 2021 at 9:41

answered Dec 9, 2021 at 13:44

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

Jonathan Hay Over a year ago

works perfect, can you help me understand the dictionary comprehension? For example, what does "i, g in df.groupby('Group')" do? How does python know what is i and what is g? and I'm use to putting an aggregator at the end of the groupby.

Jonathan Hay Over a year ago

This is a dictionary of dataframes and I'm looking for a dictionary of series. How do I move the first column over to the index within the comprehension?

jezrael Over a year ago

@JonathanHay - You are right, description was missing. Added to answer also for generate Series like need.

Jonathan Hay Over a year ago

How would this be modified to handle an additional series that only had one element (eg Beef is turned into an integer and lost)? For example df = pd.DataFrame({'col_1': ['Animal','Deer','Sheep','Fish','Vehicle','Truck','Car','Mineral','Gold','Silver','Meat',Beef], 'col_2': ['Animal',0.5,0.25,0.25,'Vehicle',0.5,0.5,'Mineral',0.75,0.25,'Meat',1], 'Group': [0,0,0,0,1,1,1,2,2,2,3,3]})

jezrael Over a year ago

@JonathanHay - I have to go away, sorry. Can you post new question with link to this solution? I have full weekend, so in Monday can check it.

Collectives™ on Stack Overflow

How to create a dictionary of series with an index from a dataframe in python

1 Answer 1

5 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

5 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related