Skip to main content
 首页 » 编程设计

python之正确地将自己的功能应用于分组的 Pandas 数据框

2025年01月19日10EasonJim

我有一个 Pandas 数据框,例如:

   ticket date         close   
0    AAA  2018-01-12  176.16 
1    AAA  2018-01-13  176.49 
3    AAA  2018-01-14  176.00 
4    BBB  2018-01-12  78.19 
5    BBB  2018-01-13  79.90 
6    BBB  2018-01-14  78.10 

我有一个函数:

def rsi(dataframe, period, column = 'close'): 
    delta = dataframe[column].diff() 
    up, down = delta.copy(), delta.copy() 
    up[up < 0] = 0 
    down[down > 0] = 0 
    rolling_up = up.ewm(com=period - 1, adjust=False).mean() 
    rolling_down = down.ewm(com= period -1, adjust=False).mean().abs() 
    rsi = 100 - 100 / (1 + rolling_up / rolling_down) 
    dataframe['rsi'] =  rsi 
    return dataframe 

我需要为每个 groupby('ticket') 将此函数应用于我的数据框。我试过了,但没用。请给我一些建议。

print(dataframe.groupby('ticket').apply(rsi, 2)) 

我得到一个错误:

cannot reindex from a duplicate axis

整个源代码是:

# -*- coding: utf-8 -*- 
 
import json 
import pandas 
import requests 
import datetime 
 
def get_historical_prices(tickets, range): 
    request_params = {'symbols': ','.join(tickets), 'types': 'chart', 'range': range} 
    json = requests.get('https://api.iextrading.com/1.0/stock/market/batch', params = request_params).json() 
    united_dataframe = pandas.DataFrame() 
    for symbol in json: 
        ticket_dataframe = pandas.DataFrame(json[symbol]['chart']) 
        ticket_dataframe.insert(0, 'ticket', symbol) 
        united_dataframe = united_dataframe.append(ticket_dataframe) 
    return united_dataframe[['ticket', 'date', 'close']] 
 
def rsi(dataframe, period, column = 'close'): 
    delta = all_prices[column].diff() 
    up, down = delta.copy(), delta.copy() 
    up[up < 0] = 0 
    down[down > 0] = 0 
    rolling_up = up.ewm(com=period - 1, adjust=False).mean() 
    rolling_down = down.ewm(com= period -1, adjust=False).mean().abs() 
    rsi = 100 - 100 / (1 + rolling_up / rolling_down) 
    dataframe['rsi'] =  rsi 
    return dataframe 
 
# Get the data 
tickets = ['AAPL', 'FB', 'TSLA'] 
all_prices = get_historical_prices(tickets, '1m') 
 
print(all_prices.groupby('ticket').apply(rsi, 2)) 

请您参考如下方法:

源代码有问题。这条线

delta = all_prices[column].diff() 

应该是

delta = dataframe[column].diff()  

修复它也可以正常运行。重新分配会将列 rsi 添加到 all_prices 即

all_prices = all_prices.groupby('ticket').apply(rsi, 2) 

所以最终的cod和结果如下图

In [20]: # -*- coding: utf-8 -*- 
    ...:  
    ...: import json 
    ...: import pandas 
    ...: import requests 
    ...: import datetime 
    ...:  
    ...: def get_historical_prices(tickets, range): 
    ...:     request_params = {'symbols': ','.join(tickets), 'types': 'chart', 'range': range} 
    ...:     json = requests.get('https://api.iextrading.com/1.0/stock/market/batch', params = request_params).json() 
    ...:     united_dataframe = pandas.DataFrame() 
    ...:     for symbol in json: 
    ...:         ticket_dataframe = pandas.DataFrame(json[symbol]['chart']) 
    ...:         ticket_dataframe.insert(0, 'ticket', symbol) 
    ...:         united_dataframe = united_dataframe.append(ticket_dataframe) 
    ...:     return united_dataframe[['ticket', 'date', 'close']] 
    ...:  
    ...: def rsi(dataframe, period, column = 'close'): 
    ...:     delta = dataframe[column].diff() 
    ...:     up, down = delta.copy(), delta.copy() 
    ...:     up[up < 0] = 0 
    ...:     down[down > 0] = 0 
    ...:     rolling_up = up.ewm(com=period - 1, adjust=False).mean() 
    ...:     rolling_down = down.ewm(com= period -1, adjust=False).mean().abs() 
    ...:     rsi = 100 - 100 / (1 + rolling_up / rolling_down) 
    ...:     dataframe['rsi'] = rsi 
    ...:     return dataframe 
    ...:  
    ...: # Get the data 
    ...: tickets = ['AAPL', 'FB', 'TSLA'] 
    ...: all_prices = get_historical_prices(tickets, '1m') 
    ...:  
    ...: all_prices = all_prices.groupby('ticket').apply(rsi, 2) 
    ...: print(all_prices.head()) 
    ...:  
    ...:  
  ticket        date   close        rsi 
0   AAPL  2018-01-12  177.09        NaN 
1   AAPL  2018-01-16  176.19   0.000000 
2   AAPL  2018-01-17  179.10  76.377953 
3   AAPL  2018-01-18  179.26  78.208232 
4   AAPL  2018-01-19  178.46  44.065484