user2242044 user2242044 - 9 months ago 62
Python Question

Pandas groupby to get two aggregated functions then convert to list of list

I'm looking for a better way to write this. This works fine for my sample data set, but is pretty slow on a larger data set. Starting with a

of customer purchases numbers. I'd like to get a list of lists in the format: ``[Customer, Mean of Orders, Count of Orders]`

import pandas as pd

df = pd.DataFrame(data = [['Customer0', 10], ['Customer0', 12], ['Customer1', 23]],
columns=['Customer', 'Orders'])

grouped = df.groupby(['Customer']).mean()
grouped['count'] = df.groupby(['Customer']).count()

values = grouped.values.tolist()
indexes = grouped.index.tolist()

for x in range(0,len(values)):
values[x].insert(0, indexes[x])

print values


[['Customer0', 11, 2], ['Customer0', 23, 1]]


Can you try this one?

df.groupby('Customer').agg(['mean', 'count']).reset_index().values.tolist()
Out: [['Customer0', 11, 2], ['Customer1', 23, 1]]

A small note: This can only improve your code significantly if the number of groups (len(values)) is quite large because we are not looping here. If you have only a small number of groups, I guess the improvement would be 2x at most.