使用另一個數據幀中的行更新數據幀

Question

如果您幫我解決這個問題，請提前致謝。我試圖完成的是使用另一個數據幀（indexed_orders）在同一日期更新一個填充了零的數據框和一個日期時間索引（我的交易數據框）。我的代碼如下：

import pandas as pd 
import numpy as np 
import os 
import csv 


orders = pd.read_csv('./orders/orders.csv', parse_dates=True, sep=',', dayfirst=True) #initiate orders data frame from csv data file 
indexed_orders = orders.set_index(['Date']) #set Date as index for orders 
print indexed_orders 

symbol_list = orders['Symbol'].tolist() #creates list of symbols 
symbols = list(set(symbol_list)) #gets rid of duplicates in list 


dates_list = orders['Date'].tolist() #creates list of order dates 
dates_orders = list(set(dates_list)) #gets rid of duplicates in list 


start_date = '2011-01-05' #establish date range 
end_date = '2011-01-20' 

dates = pd.date_range(start_date, end_date) #establish dates from start_date and end_date 

trade = pd.DataFrame(0, index = dates, columns = symbols) #establish trade data frame 
trade['Cash'] = 0 #add column for future calculations 
print trade 

這對於indexed_orders輸出：

Date   Symbol Order Shares 
2011-01-10 AAPL BUY 1500 
2011-01-13 AAPL SELL 1500 
2011-01-13 IBM BUY 4000 
2011-01-26 GOOG BUY 1000 
2011-02-02 XOM SELL 4000 
2011-02-10 XOM BUY 4000 
2011-03-03 GOOG SELL 1000 
2011-03-03 IBM SELL 2200 
2011-06-03 IBM SELL 3300 
2011-05-03 IBM BUY 1500 
2011-06-10 AAPL BUY 1200 
2011-08-01 GOOG BUY  55 
2011-08-01 GOOG SELL  55 
2011-12-20 AAPL SELL 1200 

和產出的行業如下：

  GOOG AAPL XOM IBM Cash 
2011-01-05  0  0 0 0  0 
2011-01-06  0  0 0 0  0 
2011-01-07  0  0 0 0  0 
2011-01-08  0  0 0 0  0 
2011-01-09  0  0 0 0  0 
2011-01-10  0  0 0 0  0 
2011-01-11  0  0 0 0  0 
2011-01-12  0  0 0 0  0 
2011-01-13  0  0 0 0  0 
2011-01-14  0  0 0 0  0 
2011-01-15  0  0 0 0  0 
2011-01-16  0  0 0 0  0 
2011-01-17  0  0 0 0  0 
2011-01-18  0  0 0 0  0 
2011-01-19  0  0 0 0  0 
2011-01-20  0  0 0 0  0 

我要更新目前日期我的交易數據幀在我的idexed_orders中，在正確的'Symbol'（這是交易中的AAPL，IBM，GOOG和XOM名稱）下的列中插入'Shares'數量。當indexed_orders中的'Order'列指定'SELL'時，我也希望'Shares'的值爲負值。換句話說，我想拿出代碼更新的貿易數據幀這樣的： 打印貿易

  GOOG AAPL XOM IBM Cash 
2011-01-05  0  0 0 0  0 
2011-01-06  0  0 0 0  0 
2011-01-07  0  0 0 0  0 
2011-01-08  0  0 0 0  0 
2011-01-09  0  0 0 0  0 
2011-01-10  0 1500 0 0  0 
2011-01-11  0  0 0 0  0 
2011-01-12  0  0 0 0  0 
2011-01-13  0 -1500 0 4000  0 
2011-01-14  0  0 0 0  0 
2011-01-15  0  0 0 0  0 
2011-01-16  0  0 0 0  0 
2011-01-17  0  0 0 0  0 
2011-01-18  0  0 0 0  0 
2011-01-19  0  0 0 0  0 
2011-01-20  0  0 0 0  0 

我想到某種與嵌套布爾語句迭代是必要的，但我肯定有很難找出答案。特別是，我很難想出一個方法來整合行並根據索引日期時間進行更新。

任何幫助將非常感激。

Answer 1

首先，您可以使用Order列來簽署股份變更。然後，您可以按Date和Symbol進行分組，並按求和命令進行合計。這會給你一個Series的所有獨特日子的訂單和Symbols那些日子交易。最後，使用unstack將Series轉換爲表格格式。

import numpy as np 
import pandas as pd 

df = pd.io.parsers.read_csv('temp.txt', sep = '\t') 

print df 

''' 
     Date Symbol Order Shares 
0 1/10/11 AAPL BUY 1500 
1 1/13/11 AAPL SELL 1500 
2 1/13/11 IBM BUY 4000 
3 1/26/11 GOOG BUY 1000 
4  2/2/11 XOM SELL 4000 
5 2/10/11 XOM BUY 4000 
6  3/3/11 GOOG SELL 1000 
7  3/3/11 IBM SELL 2200 
8  6/3/11 IBM SELL 3300 
9  5/3/11 IBM BUY 1500 
10 6/10/11 AAPL BUY 1200 
11 8/1/11 GOOG BUY  55 
12 8/1/11 GOOG SELL  55 
13 12/20/11 AAPL SELL 1200 
''' 

df['SharesChange'] = df.Shares * df.Order.apply(lambda o: 1 if o == 'BUY' else -1) 

df = df.groupby(['Date', 'Symbol']).agg({'SharesChange' : np.sum}).unstack().fillna(0) 

print df 
''' 
     SharesChange 
Symbol   AAPL GOOG IBM XOM 
Date 
1/10/11   1500  0  0  0 
1/13/11   -1500  0 4000  0 
1/26/11    0 1000  0  0 
12/20/11  -1200  0  0  0 
2/10/11    0  0  0 4000 
2/2/11    0  0  0 -4000 
3/3/11    0 -1000 -2200  0 
5/3/11    0  0 1500  0 
6/10/11   1200  0  0  0 
6/3/11    0  0 -3300  0 
8/1/11    0  0  0  0 
'''