Looping Through Segments in Columns then Summing

Question

action  | indicator | channel | greenfield | trucking | studio
-------------------------------------------------------------
approved| cash      | DM      | .0067      | .2356    | .1451
approved| cash      | DM      | .1067      | .4549    | .4881
approved| card      | EM      | .2067      | .8424    | .0264
approved| card      | EM      | .3067      | .5949    | .1564
approved| online    | RM      | .4555      | .4987    | .5616
...

Hello! From the table above, I've been trying to loop through each column then add based on channel, indicator, and action with no avail. Here is a manual code I've been using in it's place:

data.loc[(data['action'] == "approved") & (data['indicator'] == "cash") & 
    (data['channel'] == "DM"), 'greenfield '].sum()

The end result should look like (I am creating another table, hence the column name changes):

Segment | Name   | greenfield | trucking | studio
------------------------------------------------
DM      |Approved| .1134      | .6905    | .6332
EM      |Approved|.....       |.....     |.....

Any help in turning the above into a loop would be greatly appreciated!

Are you looking for three different sums (one for channel, indicator, action) or one sum based on the unique combinations of those columns? — Will.Evo, Aug 11 '22 at 17:28
@Will.Evo sorry for the confusion! I added a table of what I was looking for. — Nanu, Aug 11 '22 at 17:53
Loops are not what pandas was designed for. Does `groupby` solve your problem? — Vladimir Fokow, Aug 11 '22 at 17:58
@VladimirFokow unfortunately no, I've tried that. I was hoping to accomplish this with for loops. If it's not possible in Pandas, then that is the answer to my question — Nanu, Aug 11 '22 at 18:05
@VladimirFokow you are correct! when I tried groupby the first time, I di not use "sum". Thank you! — Nanu, Aug 11 '22 at 19:22

Vladimir Fokow · Accepted Answer · 2022-08-11T18:56:40.767

.group_by seems to be working:

# Constructing the dataframe:

d = {'action': ['approved', 'approved', 'approved', 'approved', 'approved'],
     'indicator': ['cash', 'cash', 'card', 'card', 'online'],
     'channel': ['DM', 'DM', 'EM', 'EM', 'RM'],
     'greenfield': [0.0067, 0.1067, 0.2067, 0.3067, 0.4555],
     'trucking': [0.2356, 0.4549, 0.8424, 0.5949, 0.4987],
     'studio': [0.1451, 0.4881, 0.0264, 0.1564, 0.5616]}

data = pd.DataFrame(d)

data.groupby(['action', 'indicator', 'channel']).sum().reset_index()

	action	indicator	channel	greenfield	trucking	studio
0	approved	card	EM	0.5134	1.4373	0.1828
1	approved	cash	DM	0.1134	0.6905	0.6332
2	approved	online	RM	0.4555	0.4987	0.5616

Looping Through Segments in Columns then Summing

1 Answers1