完整技術(shù)教程見這里

有很多人建立了自己本地的行情數(shù)據(jù)庫，希望能夠從本地數(shù)據(jù)庫將數(shù)據(jù)發(fā)到backtrader，供策略使用。一個通用的方法是將數(shù)據(jù)庫的行情數(shù)據(jù)讀到pandas dataframe里，然后將這個數(shù)據(jù)幀的數(shù)據(jù)傳給backtrader的pandas feed數(shù)據(jù)對象，這樣策略就能夠使用了。

但是有些同學(xué)不想通過pandas dataframe中轉(zhuǎn)，而是想直接從數(shù)據(jù)庫將數(shù)據(jù)喂給backtrader的數(shù)據(jù)饋送對象，這就需要針對數(shù)據(jù)庫開發(fā)專門的data feed類了。

以下就是backtrader社區(qū)提供的一個從MySQL數(shù)據(jù)庫讀取數(shù)據(jù)的feed類MySQLData，大家可以試一試。

from __future__ import (absolute_import, division, print_function,
                        unicode_literals)
import datetime
from backtrader.feed import DataBase
from backtrader import date2num
from sqlalchemy import create_engine
class MySQLData(DataBase):
    params = (
        ('dbHost', None),
        ('dbUser', None),
        ('dbPWD', None),
        ('dbName', None),
        ('ticker', 'ISL'),
        ('fromdate', datetime.datetime.min),
        ('todate', datetime.datetime.max),
        ('name', ''),
        )
    def __init__(self):
        self.engine = create_engine('mysql://'+self.p.dbUser+':'+ self.p.dbPWD +'@'+ self.p.dbHost +'/'+ self.p.dbName +'?charset=utf8mb4', echo=False)
    def start(self):
        self.conn = self.engine.connect()
        self.stockdata = self.conn.execute('SELECT id FROM stocks WHERE ticker LIKE '' + self.p.ticker + '' LIMIT 1')
        self.stock_id = self.stockdata.fetchone()[0]
        #self.result = self.engine.execute('SELECT `date`,`open`,`high`,`low`,`close`,`volume` FROM `eoddata` WHERE `stock_id` = 10 AND `date` between ''+self.p.fromdate.strftime('%Y-%m-%d')+'' and ''+self.p.todate.strftime('%Y-%m-%d')+'' ORDER BY `date` ASC')
        self.result = self.conn.execute('SELECT `date`,`open`,`high`,`low`,`close`,`volume` FROM `eoddata` WHERE `stock_id` = ' + str(self.stock_id) + ' AND `date` between ''+self.p.fromdate.strftime('%Y-%m-%d')+'' and ''+self.p.todate.strftime('%Y-%m-%d')+'' ORDER BY `date` ASC')
    def stop(self):
        #self.conn.close()
        self.engine.dispose()
    def _load(self):
        one_row = self.result.fetchone()
        if one_row is None:
            return False
        self.lines.datetime[0] = date2num(one_row[0])
        self.lines.open[0] = float(one_row[1])
        self.lines.high[0] = float(one_row[2])
        self.lines.low[0] = float(one_row[3])
        self.lines.close[0] = float(one_row[4])
        self.lines.volume[0] = int(one_row[5])
        self.lines.openinterest[0] = -1
        return True

發(fā)布于 19 小時前