5-1. groupby1(groupby, swaplevel, unstack, aggregation, transformation, filteration)

df.groupby된 정보는 object에 불과함

grouped = df.groupby("Teams")

grouped

Out[17]:

<pandas.core.groupby.groupby.DataFrameGroupBy object at 0x0000028890B2D710>

In [19]:

for data in grouped:

    print(data)

('devils',     Teams  Rank  Year  Points
2  devils     2    14     136
3  devils     3    15     123)
('kings',    Teams  Rank  Year  Points
4  kings     3    14     878
5  kings     4    15     533
6  kings     1    16     234
7  kings     1    17     345)
('riders',      Teams  Rank  Year  Points
0   riders     1    14     845
1   riders     2    15     734
8   riders     2    16     745
11  riders     2    17     463)
('royals',      Teams  Rank  Year  Points
9   royals     4    14     113
10  royals     1    15     345)

df.groupby(["column1", "column2"])["data"].연산함수()

➡컬럼1과 컬럼2를 기준으로 컬럼을 그룹화 시킨다음에, 그 안에 들어가는 실질적인 데이터는 data에 연산함수 적용한걸 넣을거야

grouped = df.groupby("Teams")

grouped

Out[17]:

<pandas.core.groupby.groupby.DataFrameGroupBy object at 0x0000028890B2D710>

In [19]:

for data in grouped:

    print(data)

('devils',     Teams  Rank  Year  Points
2  devils     2    14     136
3  devils     3    15     123)
('kings',    Teams  Rank  Year  Points
4  kings     3    14     878
5  kings     4    15     533
6  kings     1    16     234
7  kings     1    17     345)
('riders',      Teams  Rank  Year  Points
0   riders     1    14     845
1   riders     2    15     734
8   riders     2    16     745
11  riders     2    17     463)
('royals',      Teams  Rank  Year  Points
9   royals     4    14     113 

10 royals 1 15 345)