Pandas中concat连接多个DataFrame

import pandas
# 读取stocks_2016和stocks_2017两个数据集，用Symbol作为行索引名
stocks_2016 = pd.read_csv('data/stocks_2016.csv', index_col='Symbol')
stocks_2017 = pd.read_csv('data/stocks_2017.csv', index_col='Symbol')
stocks_2016

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

stocks_2017

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

将两个DataFrame放到一个列表中，用pandas的concat方法将它们连接起来

s_list = [stocks_2016, stocks_2017]
pd.concat(s_list)

	Shares	Low	High
Symbol
AAPL	80	95	110
TSLA	50	80	130
WMT	40	55	70
AAPL	50	120	140
GE	100	30	40
IBM	87	75	95
SLB	20	55	85
TXN	500	15	23
TSLA	100	100	300

# keys参数可以给两个DataFrame命名，该标签会出现在行索引的最外层，
会生成多层索引(multiIndex)，names参数可以重命名每个索引层
pd.concat(s_list, keys=['2016', '2017'], names=['Year', 'Symbol'])

# keys参数可以给两个DataFrame命名，该标签会出现在行索引的最外层，会生成多层索引(multiIndex)，names参数可以重命名每个索引层
pd.concat(s_list, keys=['2016', '2017'], names=['Year', 'Symbol'])

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

# 也可以横向连接。只要将axis参数设为columns或1
pd.concat(s_list, keys=['2016', '2017'], axis='columns', names=['Year', None])

Year	2016			2017
	Shares	Low	High	Shares	Low	High
AAPL	80.0	95.0	110.0	50.0	120.0	140.0
TSLA	50.0	80.0	130.0	100.0	100.0	300.0
WMT	40.0	55.0	70.0	NaN	NaN	NaN
GE	NaN	NaN	NaN	100.0	30.0	40.0
IBM	NaN	NaN	NaN	87.0	75.0	95.0
SLB	NaN	NaN	NaN	20.0	55.0	85.0
TXN	NaN	NaN	NaN	500.0	15.0	23.0

# concat函数默认使用的是外连接，会保留每个DataFrame中的所有行。也可以通过设定join参数，使用内连接：
pd.concat(s_list, join='inner', keys=['2016', '2017'], axis='columns', names=['Year', None])

Year	2016			2017
	Shares	Low	High	Shares	Low	High
Symbol
AAPL	80	95	110	50	120	140
TSLA	50	80	130	100	100	300

append拼接DataFrame

# append是concat方法的超简化版本，append内部其实就是调用concat。pd.concat也可以
stocks_2016.append(stocks_2017)

	Shares	Low	High
Symbol
AAPL	80	95	110
TSLA	50	80	130
WMT	40	55	70
AAPL	50	120	140
GE	100	30	40
IBM	87	75	95
SLB	20	55	85
TXN	500	15	23
TSLA	100	100	300

#copy() 函数返回一个字典的浅复制。
stocks_2015 = stocks_2016.copy() #Python 字典(Dictionary) copy() 函数返回一个字典的浅复制。浅复制是指当对象的字段值被复制时，字段引用的对象不会被复制
stocks_2015

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

Pandas中concat连接多个DataFrame

猜你喜欢