股票涨跌预测方法之一：下载股票数据

         前一阵子在同学的鼓动下，花了一个多月研究了股票行情的预测方法，熟悉了常见的炒股术语及技术指标，现总结如下，纯属兴趣，如果想依照本文的方法来短线操作获利，请绕道。

         研究的第一步就是下载所有股票的历史行情数据。可以使用python开源模块tushare，缺点就是慢，所以我使用多线程来加速，并将数据保存在sqlite数据库里，下面是python代码：

# -*- coding: utf-8 -*-

import tushare as ts
import pickle
import sqlite3 as db
import datetime
from queue import *
import threading

g_taskList = Queue(0)
g_resList = Queue(0)

def task():
    while True:
        try:
            id, y1, y2 = g_taskList.get(False)
        except:
            break
            
        try:
            record = []
            flag = 0
            for year in range(y2,y1-1,-1):
                if flag>=2: break
                v1 = ts.get_h_data(id,start='%d-06-01'%year,end='%d-12-31'%year) 
                if v1 is None:
                    flag += 1
                    continue
                index = [m.date() for m in v1.index]
                for v in zip(index, v1.open, v1.high, v1.close, v1.low, v1.volume, v1.amount):
                    record.append([id,]+list(v))
            g_resList.put([id, record])            
        except Exception as e:
            g_taskList.put([id, y1, y2])
            break
    g_resList.put([None,[]])


def GetAllData(dbFile='all_tushare_data2017.db', y1= 1990, y2=2017):
    
    data = list(ts.get_stock_basics().index)                                            #这里可以获取股票代码 
    try:
        cxn = db.connect(dbFile)
        cur = cxn.cursor()
        
        cur.execute('create table if not exists gp_record(code char(6), date date, open float, high float, close float, low float, volume float, amout float, primary key  (code, date))')   #创建表格
    
        for d in data:  
            g_taskList.put([d,y1,y2])
            
        count = 20                                                  #启动20个下载线程
        for k in range(count):
            threading.Thread(None, task).start()
        
        while True:
            if count<=0:        break
            id, record = g_resList.get()
            if not id:
                count-=1; continue
            print(id, len(record))
            for v in record:
                try:
                    cur.execute('insert into gp_record values(?,?,?,?,?,?,?,?)', v)
                except:
                    #print('insert error',d)
                    pass
            cxn.commit()
    finally:
        cur.close()
        cxn.commit()
        cxn.close()
    
if __name__ == '__main__':
    GetAllData()