(11주차) 11월17일

• 강의영상
• import
• data
• pd.concat([...])
  • 예제1
  • 예제2
• pd.concat([...],axis='columns')
  • 예제3
• pd.concat([...],keys=[...])
  • 예제4
• pd.concat([...],keys=[...],axis='columns')
  • 예제5
• 숙제

강의영상

- (1/3) 데이터프레임을 합치는 방법 (1)

- (2/3) 데이터프레임을 합치는 방법 (2)

- (3/3) 데이터프레임을 합치는 방법 (3), zip 활용방법

import

import pandas as pd 

data

df2016=pd.read_csv("https://raw.githubusercontent.com/PacktPublishing/Pandas-Cookbook/master/data/stocks_2016.csv")
df2017=pd.read_csv("https://raw.githubusercontent.com/PacktPublishing/Pandas-Cookbook/master/data/stocks_2017.csv")
df2018=pd.read_csv("https://raw.githubusercontent.com/PacktPublishing/Pandas-Cookbook/master/data/stocks_2018.csv")

pd.concat([...])

예제1

df2016

	Symbol	Shares	Low	High
0	AAPL	80	95	110
1	TSLA	50	80	130
2	WMT	40	55	70

df2017

	Symbol	Shares	Low	High
0	AAPL	50	120	140
1	GE	100	30	40
2	IBM	87	75	95
3	SLB	20	55	85
4	TXN	500	15	23
5	TSLA	100	100	300

df2018

	Symbol	Shares	Low	High
0	AAPL	40	135	170
1	AMZN	8	900	1125
2	TSLA	50	220	400

pd.concat([df2016,df2017,df2018]) 

	Symbol	Shares	Low	High
0	AAPL	80	95	110
1	TSLA	50	80	130
2	WMT	40	55	70
0	AAPL	50	120	140
1	GE	100	30	40
2	IBM	87	75	95
3	SLB	20	55	85
4	TXN	500	15	23
5	TSLA	100	100	300
0	AAPL	40	135	170
1	AMZN	8	900	1125
2	TSLA	50	220	400

pd.concat([df2016,df2017,df2018]).reset_index()

	index	Symbol	Shares	Low	High
0	0	AAPL	80	95	110
1	1	TSLA	50	80	130
2	2	WMT	40	55	70
3	0	AAPL	50	120	140
4	1	GE	100	30	40
5	2	IBM	87	75	95
6	3	SLB	20	55	85
7	4	TXN	500	15	23
8	5	TSLA	100	100	300
9	0	AAPL	40	135	170
10	1	AMZN	8	900	1125
11	2	TSLA	50	220	400

pd.concat([df2016,df2017,df2018]).reset_index(drop=True)

	Symbol	Shares	Low	High
0	AAPL	80	95	110
1	TSLA	50	80	130
2	WMT	40	55	70
3	AAPL	50	120	140
4	GE	100	30	40
5	IBM	87	75	95
6	SLB	20	55	85
7	TXN	500	15	23
8	TSLA	100	100	300
9	AAPL	40	135	170
10	AMZN	8	900	1125
11	TSLA	50	220	400

- R에서 rbind와 유사하다

예제2

- rbind와의 차이점: 합치려는 데이터 프레임의 columns이 꼭 동일할 필요는 없다.

df2016

	Symbol	Shares	Low	High
0	AAPL	80	95	110
1	TSLA	50	80	130
2	WMT	40	55	70

df2017.iloc[:,1:]

	Shares	Low	High
0	50	120	140
1	100	30	40
2	87	75	95
3	20	55	85
4	500	15	23
5	100	100	300

pd.concat([df2016,df2017.iloc[:,1:]])

	Symbol	Shares	Low	High
0	AAPL	80	95	110
1	TSLA	50	80	130
2	WMT	40	55	70
0	NaN	50	120	140
1	NaN	100	30	40
2	NaN	87	75	95
3	NaN	20	55	85
4	NaN	500	15	23
5	NaN	100	100	300

pd.concat([...],axis='columns')

- R에서 cbind와 비슷한 느낌이다. (그런데 row의 숫자가 서로 달라도 괜찮음)

예제3

df2016

	Symbol	Shares	Low	High
0	AAPL	80	95	110
1	TSLA	50	80	130
2	WMT	40	55	70

df2017

	Symbol	Shares	Low	High
0	AAPL	50	120	140
1	GE	100	30	40
2	IBM	87	75	95
3	SLB	20	55	85
4	TXN	500	15	23
5	TSLA	100	100	300

df2018

	Symbol	Shares	Low	High
0	AAPL	40	135	170
1	AMZN	8	900	1125
2	TSLA	50	220	400

pd.concat([df2016,df2017,df2018],axis='columns')

	Symbol	Shares	Low	High	Symbol	Shares	Low	High	Symbol	Shares	Low	High
0	AAPL	80.0	95.0	110.0	AAPL	50	120	140	AAPL	40.0	135.0	170.0
1	TSLA	50.0	80.0	130.0	GE	100	30	40	AMZN	8.0	900.0	1125.0
2	WMT	40.0	55.0	70.0	IBM	87	75	95	TSLA	50.0	220.0	400.0
3	NaN	NaN	NaN	NaN	SLB	20	55	85	NaN	NaN	NaN	NaN
4	NaN	NaN	NaN	NaN	TXN	500	15	23	NaN	NaN	NaN	NaN
5	NaN	NaN	NaN	NaN	TSLA	100	100	300	NaN	NaN	NaN	NaN

pd.concat([df2016,df2017,df2018],axis=1)

	Symbol	Shares	Low	High	Symbol	Shares	Low	High	Symbol	Shares	Low	High
0	AAPL	80.0	95.0	110.0	AAPL	50	120	140	AAPL	40.0	135.0	170.0
1	TSLA	50.0	80.0	130.0	GE	100	30	40	AMZN	8.0	900.0	1125.0
2	WMT	40.0	55.0	70.0	IBM	87	75	95	TSLA	50.0	220.0	400.0
3	NaN	NaN	NaN	NaN	SLB	20	55	85	NaN	NaN	NaN	NaN
4	NaN	NaN	NaN	NaN	TXN	500	15	23	NaN	NaN	NaN	NaN
5	NaN	NaN	NaN	NaN	TSLA	100	100	300	NaN	NaN	NaN	NaN

• 칼럼이 길잖아요.. 그래서 1이에요..

pd.concat([...],keys=[...])

예제4

df2016

	Symbol	Shares	Low	High
0	AAPL	80	95	110
1	TSLA	50	80	130
2	WMT	40	55	70

df2017

	Symbol	Shares	Low	High
0	AAPL	50	120	140
1	GE	100	30	40
2	IBM	87	75	95
3	SLB	20	55	85
4	TXN	500	15	23
5	TSLA	100	100	300

df2018

	Symbol	Shares	Low	High
0	AAPL	40	135	170
1	AMZN	8	900	1125
2	TSLA	50	220	400

pd.concat([df2016,df2017,df2018],keys=[2016,2017,2018])

		Symbol	Shares	Low	High
2016	0	AAPL	80	95	110
	1	TSLA	50	80	130
	2	WMT	40	55	70
2017	0	AAPL	50	120	140
	1	GE	100	30	40
	2	IBM	87	75	95
	3	SLB	20	55	85
	4	TXN	500	15	23
	5	TSLA	100	100	300
2018	0	AAPL	40	135	170
	1	AMZN	8	900	1125
	2	TSLA	50	220	400

- 위의 코드는 아래와 같다.

pd.concat({2016:df2016,2017:df2017,2018:df2018})

		Symbol	Shares	Low	High
2016	0	AAPL	80	95	110
	1	TSLA	50	80	130
	2	WMT	40	55	70
2017	0	AAPL	50	120	140
	1	GE	100	30	40
	2	IBM	87	75	95
	3	SLB	20	55	85
	4	TXN	500	15	23
	5	TSLA	100	100	300
2018	0	AAPL	40	135	170
	1	AMZN	8	900	1125
	2	TSLA	50	220	400

- 그런데 {2016:df2016,2017:df2017,2018:df2018} 와 dict(zip([2016,2017,2018], [df2016,df2017,df2018])) 는 같으므로 위의코드는 다시 아래와 같다.

pd.concat(dict(zip([2016,2017,2018],[df2016,df2017,df2018])))

		Symbol	Shares	Low	High
2016	0	AAPL	80	95	110
	1	TSLA	50	80	130
	2	WMT	40	55	70
2017	0	AAPL	50	120	140
	1	GE	100	30	40
	2	IBM	87	75	95
	3	SLB	20	55	85
	4	TXN	500	15	23
	5	TSLA	100	100	300
2018	0	AAPL	40	135	170
	1	AMZN	8	900	1125
	2	TSLA	50	220	400

- 의도하지 않은 코드들

pd.concat([df2016,df2017,df2018],keys=[2016,2017])

		Symbol	Shares	Low	High
2016	0	AAPL	80	95	110
	1	TSLA	50	80	130
	2	WMT	40	55	70
2017	0	AAPL	50	120	140
	1	GE	100	30	40
	2	IBM	87	75	95
	3	SLB	20	55	85
	4	TXN	500	15	23
	5	TSLA	100	100	300

pd.concat([df2016,df2017,df2018],keys=[2016,2018])

		Symbol	Shares	Low	High
2016	0	AAPL	80	95	110
	1	TSLA	50	80	130
	2	WMT	40	55	70
2018	0	AAPL	50	120	140
	1	GE	100	30	40
	2	IBM	87	75	95
	3	SLB	20	55	85
	4	TXN	500	15	23
	5	TSLA	100	100	300

pd.concat([df2016,df2017,df2018],keys=[2016,2017,2018,2019])

		Symbol	Shares	Low	High
2016	0	AAPL	80	95	110
	1	TSLA	50	80	130
	2	WMT	40	55	70
2017	0	AAPL	50	120	140
	1	GE	100	30	40
	2	IBM	87	75	95
	3	SLB	20	55	85
	4	TXN	500	15	23
	5	TSLA	100	100	300
2018	0	AAPL	40	135	170
	1	AMZN	8	900	1125
	2	TSLA	50	220	400

pd.concat([...],keys=[...],axis='columns')

예제5

df2016

	Symbol	Shares	Low	High
0	AAPL	80	95	110
1	TSLA	50	80	130
2	WMT	40	55	70

df2017

	Symbol	Shares	Low	High
0	AAPL	50	120	140
1	GE	100	30	40
2	IBM	87	75	95
3	SLB	20	55	85
4	TXN	500	15	23
5	TSLA	100	100	300

df2018

	Symbol	Shares	Low	High
0	AAPL	40	135	170
1	AMZN	8	900	1125
2	TSLA	50	220	400

pd.concat([df2016,df2017,df2018],axis=1)

	Symbol	Shares	Low	High	Symbol	Shares	Low	High	Symbol	Shares	Low	High
0	AAPL	80.0	95.0	110.0	AAPL	50	120	140	AAPL	40.0	135.0	170.0
1	TSLA	50.0	80.0	130.0	GE	100	30	40	AMZN	8.0	900.0	1125.0
2	WMT	40.0	55.0	70.0	IBM	87	75	95	TSLA	50.0	220.0	400.0
3	NaN	NaN	NaN	NaN	SLB	20	55	85	NaN	NaN	NaN	NaN
4	NaN	NaN	NaN	NaN	TXN	500	15	23	NaN	NaN	NaN	NaN
5	NaN	NaN	NaN	NaN	TSLA	100	100	300	NaN	NaN	NaN	NaN

pd.concat([df2016,df2017,df2018],axis=1,keys=[2016,2017,2018])

	2016				2017				2018
	Symbol	Shares	Low	High	Symbol	Shares	Low	High	Symbol	Shares	Low	High
0	AAPL	80.0	95.0	110.0	AAPL	50	120	140	AAPL	40.0	135.0	170.0
1	TSLA	50.0	80.0	130.0	GE	100	30	40	AMZN	8.0	900.0	1125.0
2	WMT	40.0	55.0	70.0	IBM	87	75	95	TSLA	50.0	220.0	400.0
3	NaN	NaN	NaN	NaN	SLB	20	55	85	NaN	NaN	NaN	NaN
4	NaN	NaN	NaN	NaN	TXN	500	15	23	NaN	NaN	NaN	NaN
5	NaN	NaN	NaN	NaN	TSLA	100	100	300	NaN	NaN	NaN	NaN

pd.concat({2016:df2016,2017:df2017,2018:df2018},axis=1)

	2016				2017				2018
	Symbol	Shares	Low	High	Symbol	Shares	Low	High	Symbol	Shares	Low	High
0	AAPL	80.0	95.0	110.0	AAPL	50	120	140	AAPL	40.0	135.0	170.0
1	TSLA	50.0	80.0	130.0	GE	100	30	40	AMZN	8.0	900.0	1125.0
2	WMT	40.0	55.0	70.0	IBM	87	75	95	TSLA	50.0	220.0	400.0
3	NaN	NaN	NaN	NaN	SLB	20	55	85	NaN	NaN	NaN	NaN
4	NaN	NaN	NaN	NaN	TXN	500	15	23	NaN	NaN	NaN	NaN
5	NaN	NaN	NaN	NaN	TSLA	100	100	300	NaN	NaN	NaN	NaN

pd.concat(dict(zip([2016,2017,2018],[df2016,df2017,df2018])),axis=1)

	2016				2017				2018
	Symbol	Shares	Low	High	Symbol	Shares	Low	High	Symbol	Shares	Low	High
0	AAPL	80.0	95.0	110.0	AAPL	50	120	140	AAPL	40.0	135.0	170.0
1	TSLA	50.0	80.0	130.0	GE	100	30	40	AMZN	8.0	900.0	1125.0
2	WMT	40.0	55.0	70.0	IBM	87	75	95	TSLA	50.0	220.0	400.0
3	NaN	NaN	NaN	NaN	SLB	20	55	85	NaN	NaN	NaN	NaN
4	NaN	NaN	NaN	NaN	TXN	500	15	23	NaN	NaN	NaN	NaN
5	NaN	NaN	NaN	NaN	TSLA	100	100	300	NaN	NaN	NaN	NaN

숙제

- 예제1~5의 결과를 타이디한 데이터로 바꿔볼 것 (제출의무는 없음)