CODEWITHSUNDEEP
pythongroup-bypandas
stephenlee
stephenlee

Reputation: 55

python pandas groupby filter

Concretely, say I have a DataFrame like this:

appid  mac_id  count

1      a        1 

2      b        1

2      c        1

3      d        1

3      e        1

And I also have a :

mac_list = ['b', 'd', 'e']

I want to group this data frame on appid and for every group filter mac_id if it's in mac_list. Last, sum(count) for every group.

for this DataFrame the result is:

appid   count

1         0

2         1

3         2

How can I do this with Pandas?

Upvotes: 1

Views: 2502

Answers (1)

Katriel
Katriel

Reputation: 123772

>>> df = pd.DataFrame({"appid": [1,2,2,3,3], "mac_id": ['a', 'b', 'c', 'd', 'e'], "count": [1,1,1,1,1]})
>>> summer = lambda x: x[x["mac_id"].isin(mac_list)].sum()
>>> df.groupby("appid").apply(summer)["count"]
  18 
appid
1        0
2        1
3        2
Name: count, dtype: object

Upvotes: 7

Related Questions