在线下载CSV并转换为Pandas DF

Question

我有一个在线下载的CSV文件。我正试图将它放入Pandas数据帧中。来源是联邦储备银行和使用通货膨胀数据(CPI(。我错过了一些能让它发挥作用的东西。

以下是要下载的URL：

s_url = https://fred.stlouisfed.org/graph/fredgraph.csv?cosd=1990-01-01&coed=2040-01-01&id=CPIAUCSL

我拿着那个URL，读着这个：

response = urlopen(s_url).read()

这给了我一个看起来像这样的东西：

b'DATE,CPIAUCSLn1990-01-01,127.5n1990-02-01,128.0n1990-03-01,128.6n1990-04-01,128.9n1990-05-01,129.1n1990-06-01,129.9n1990-07-01,130.5n1990-08-01,131.6n1990-09-01,132.5n1990-10-01,133.4n1990-11-01,133.7n1990-12-01,134.2n1991-01-01,134.7n1991-02-01,134.8n1991-03-01,134.8n1991-04-01,135.1n1991-05-01,135.6n1991-06-01,136.0n'

然后我试着把这个回应读到Pandas:中

df = pd.read_csv(response)

这行不通。在这个过程中我缺少了哪些步骤？

Answer 1

似乎可以直接使用url作为df参数。你试过了吗？

import pandas as pd 
s_url = 'https://fred.stlouisfed.org/graph/fredgraph.csv?cosd=1990-01-01&coed=2040-01-01&id=CPIAUCSL' 
df = pd.read_csv(s_url)

Answer 2

您可以使用stringIO从字符串中读取数据。

data = b'DATE,CPIAUCSLn1990-01-01,127.5n1990-02-01,128.0n1990-03-01,128.6n1990-04-01,128.9n1990-05-01,129.1n1990-06-01,129.9n1990-07-01,130.5n1990-08-01,131.6n1990-09-01,132.5n1990-10-01,133.4n1990-11-01,133.7n1990-12-01,134.2n1991-01-01,134.7n1991-02-01,134.8n1991-03-01,134.8n1991-04-01,135.1n1991-05-01,135.6n1991-06-01,136.0n'
from io import StringIO
import pandas as pd
data = data.decode('utf-8') #decode binary string to utf-8
df = pd.read_csv(StringIO(data)) #read the data
df

输出：

DATE    CPIAUCSL
0   1990-01-01  127.5
1   1990-02-01  128.0
2   1990-03-01  128.6
3   1990-04-01  128.9
4   1990-05-01  129.1
5   1990-06-01  129.9