用于在Python中从多个csv导入和计算股票回报的循环

Question

也许是个愚蠢的问题，但我是个初学者。我使用下面的代码来计算股票收益率和平方对数收益率：

#read csv
stock1 = pd.read_csv('http://mydata.com/q/c/?d=stock1%d1=20190101&d2=20200101')
stock2 = pd.read_csv('http://mydata.com/q/c/?d=stock2%d1=20190101&d2=20200101')

#calculate return
stock1['r'] = np.log(stock1.close) - np.log(stock1.close.shift(1))
stock2['r'] = np.log(stock2.close) - np.log(stock2.close.shift(1))

#calculate squared returns and rolling returns
stock1_sq_r = stock1['r'] * stock1['r']
stock2_sq_r = stock2['r'] * stock2['r']

rolled_s1_sq_r = stock1_sq_r.rolling(window=5).sum()
rolled_s2_sq_r = stock2_sq_r.rolling(window=5).sum()

它很有魅力，但我需要为100只股票这么做。。。我怎样才能把上面的循环

我想读取csv的循环应该与此类似：

stocks = {'stock1', 'stock2', 'stock3'}
      for stock in stocks:
      url = ('http://http://mydata.com/q/c/?d={}%d1=20190101&d2=20200101')
      dfs = pd.read_csv(url.format(stock))

但它只读取stock3的数据，我不知道如何循环其余的计算

谢谢你的建议、指导和线索

Answer 1

压痕可能是您的问题。试试这个：

stocks = {'stock1', 'stock2', 'stock3'}
      for stock in stocks:
          url = ('http://http://mydata.com/q/c/?d={}%d1=20190101&d2=20200101')
          dfs = pd.read_csv(url.format(stock))

请注意for循环中的缩进

Answer 2

也许这样会更好：

stocks = ['stock1', 'stock2', 'stock3']
for stock in stocks:
   url = f'http://mydata.com/q/c/?d={stock}%d1=20190101&d2=20200101'
   dfs = pd.read_csv(url)

Answer 3

您可以将所有值读入字典（stock_name作为键，dataframe作为值）

all_data_dict ={}

stocks = {'stock1', 'stock2', 'stock3'}
for stock in stocks:
   url = 'http://http://mydata.com/q/c/?d={}%d1=20190101&d2=20200101'.format(stock)
   all_data_dict[stock] = pd.read_csv(url)

您可以通过以下方式访问字典中“stock1”的数据帧

df = all_data_dict['stock1']

您还可以使用pandas.DataFrame.diff查找与下一个值的差值&；使用**获得正方形

rolled_s1_sq_r = (np.log(df.close).diff()**2).rolling(window=5).sum()