1. 從不同數(shù)據(jù)來源獲取——本地
1.1 常用：本地文件讀取
with open('data/000001.csv', 'r') as f: #推薦這種方法；
? ? for i in range(5):
? ? ? ? print(f.readline())
1.2 Python CSV模塊讀取
import csv
csv_reader = csv.reader(open('data/000001.csv', 'r'))
1.3 常用：Pandas讀取CSV
import pandas as pd
import numpy as np
data = pd.read_csv('data/000001.csv')
data = pd.read_csv('data/000001.csv', index_col=1,
? ? ? ? ? ? ? ? ? parse_dates=True)? ? ? ? ? ? ? ? #解析日期
2. 從網(wǎng)絡(luò)Open Source讀取
2.1 Yahoo
from pandas_datareader import data as web
import fix_yahoo_finance as yf
yf.pdr_override()
#不翻墻或者網(wǎng)速較慢可能無法從Yahoo讀取，屬于正诚唬現(xiàn)象廉羔；
data = web.get_data_yahoo('GS', start = '2010-01-01', end = '2012-01-01')
data= web.get_data_yahoo('MSFT', start = '2016-01-01', end = '2017-06-30')
data= web.get_data_yahoo('600030.ss', start = '2016-01-01', end = '2017-07-01')
2.2 Quandl
import quandl
data = quandl.get('EOD/KO',start_date='2016-1-1',end_date='2017-06-30')
2.3 常用：Tushare
獲取結(jié)構(gòu)化行情數(shù)據(jù)
import pandas as pd
import tushare as ts
hs300 = ts.get_k_data('hs300',start ='2015-01-01', end = '2017-06-30') #get_k_data()
hs300.set_index('date', inplace = True) #pd.set_index(),將df中的某一列設(shè)置成為索引抄罕；
hs300.head()
hs300['close'].plot(figsize=(10, 6))
hs300.close.plot(figsize=(10, 6)) #等價锦募；
data = ts.get_k_data('600030') #默認(rèn)前復(fù)權(quán)價格虑瀑；
data2 = ts.get_k_data('600030', autype='hfq') #不復(fù)權(quán)
data3 = ts.get_k_data('600030', ktype = '5') #兩個日期之間的前復(fù)權(quán)數(shù)據(jù)
data = ts.get_k_data(['600030','000001']) #tushare API接口不支持多股票數(shù)據(jù)逛拱；
df = ts.get_tick_data('600030',date='2017-07-28') #get_tick_data()
df.sort_indexs(inplace = True, ascending = False)
Tushare獲得當(dāng)前主流指數(shù)列表
df = ts.get_index()
Tushare獲得股票的基本面信息
df = ts.get_stock_basics() #基本面數(shù)據(jù)
date = df.ix['600848']['timeToMarket']
date = df.loc['600030']['timeToMarket'] #ix即將要被取消敌厘；
獲得所有股票基本面數(shù)據(jù)
data = ts.get_stock_basics() #get_stocl_basics()
data.ix['600030'][['pe','esp']] #pandas數(shù)據(jù)選擇的復(fù)習(xí)；
data = ts.get_profit_data(2017,1) #獲得公司盈利數(shù)據(jù)朽合；
ts.get_latest_news(top=5,show_content=True) #顯示最新5條新聞俱两，并打印出新聞內(nèi)容
top_list = ts.top_list('2017-08-11')
2.4 常用：優(yōu)礦
可以通過優(yōu)礦下載數(shù)據(jù)饱狂，并保存成CSV文件下載再導(dǎo)入；
# 獲得某一只當(dāng)天的tick數(shù)據(jù)宪彩；
data=DataAPI.MktTickRTIntraDayGet(securityID=u"000001.XSHE",startTime=u"09:30",endTime=u"15:00",field=u"",pandas="1")
data.to_csv('tick_data.csv')? #下載并保存數(shù)據(jù)以供分析休讳；
# 獲得某一些股票具體某一天的因子數(shù)據(jù)；
DataAPI.MktStockFactorsOneDayGet(tradeDate=u"20170630",secID=u"",ticker=u"000001,600030",field=u"ticker,ROE,PE,PB",pandas="1")
#某一只股票一段時間之內(nèi)的因子數(shù)據(jù)尿孔；
DataAPI.MktStockFactorsDateRangeGet(secID=u"",ticker=u"000001",beginDate=u"20100101",endDate=u"20170616",field=u"tradeDate,ROE,PE,PB",pandas="1")
# 獲取交易日歷
start_date = '2014-01-01'
end_date = '2017-07-01'
trading_date = DataAPI.TradeCalGet(exchangeCD=u"XSHG",beginDate=u"",endDate=u"",field=u"",pandas="1")
# trading_date.to_csv('trading_date.csv')
# 篩選2013年到2016年每月最后一個交易日的日期
print(trading_date)
month_end = trading_date[(trading_date['isMonthEnd']==1) & (trading_date['calendarDate']>start_date) & (trading_date['calendarDate']<end_date)]['calendarDate'].tolist()
print month_end
# 獲取某個日期以前上市的俊柔，正常交易或暫停交易的股票代碼，格式為xxxxxx.XSHE或xxxxxx.XSHG
date = '2017-10-01'
stock_basics = DataAPI.EquGet(equTypeCD=u"A",secID=u"",ticker=u"",listStatusCD=u"",field=u"",pandas="1")
# stock_basics.to_csv('data/stock_basics.csv', encoding='GB18030')
valid_stocks = stock_basics.loc[(stock_basics['listDate']<date) & (stock_basics['listStatusCD'].isin(['L','S']))]['secID']
# valid_stocks.to_csv('data/valid_stocks.csv', encoding='GB18030')
print valid_stocks
# 獲取對應(yīng)股票在對應(yīng)日期的多個因子值
import pandas as pd
mkt_value = [DataAPI.MktStockFactorsOneDayGet(tradeDate=date,secID=valid_stocks,ticker=u"",field=["secID", 'LCAP','PE', 'REVS20', 'tradeDate'],pandas="1").set_index(['tradeDate', 'secID']) for date in month_end]
lcap = pd.concat(mkt_value, axis=0)
# lcap.to_csv('data/raw_factors.csv')
print lcap.head(5)
# 每個月最后一個交易日計算市值最小的20只股票
import pandas as pd
min_cap_pool = {date: lcap['LCAP'][date].sort_values(ascending=True).index[:20] for date in month_end}
min_cap_pool = pd.DataFrame(min_cap_pool)
print min_cap_pool
