Python读取通达信本地数据
一、通达信日线*.day文件
文件名即股票代码
每32个字节为一天数据
每4个字节为一个字段，每个字段内低字节在前
00 ~ 03 字节：年月日, 整型
04 ~ 07 字节：开盘价*1000， 整型
08 ~ 11 字节：最高价*1000, 整型
12 ~ 15 字节：最低价*1000, 整型
16 ~ 19 字节：收盘价*1000, 整型
20 ~ 23 字节：成交额（元），float型
24 ~ 27 字节：成交量（手），整型
28 ~ 31 字节：上日收盘*1000, 整型
二、通达信5分钟线*.5文件
每32个字节为一个5分钟数据，每字段内低字节在前
00 ~ 01 字节：日期，整型，设其值为num，则日期计算方法为：
year=floor(num/2048)+2004;
month=floor(mod(num,2048)/100);
day=mod(mod(num,2048),100);
02 ~ 03 字节： 从0点开始至目前的分钟数，整型
04 ~ 07 字节：开盘价（分），整型
08 ~ 11 字节：最高价（分），整型
12 ~ 15 字节：最低价（分），整型
16 ~ 19 字节：收盘价（分），整型
20 ~ 23 字节：成交额（元），float型
24 ~ 27 字节：成交量（股）
28 ~ 31 字节：（保留）
三、在 vipdoc/sh 目录下新建了个 pythondata 文件夹，注意文件路径分隔符是 /
import struct import datetime def stock_csv(filepath, name): data = [] with open(filepath, \'rb\') as f: file_object_path = \'D:/通达信/vipdoc/sh/pythondata/\' + name +\'.csv\' file_object = open(file_object_path, \'w+\') while True: stock_date = f.read(4) stock_open = f.read(4) stock_high = f.read(4) stock_low= f.read(4) stock_close = f.read(4) stock_amount = f.read(4) stock_vol = f.read(4) stock_reservation = f.read(4) # date,open,high,low,close,amount,vol,reservation if not stock_date: break stock_date = struct.unpack("l", stock_date) # 4字节 如20091229 stock_open = struct.unpack("l", stock_open) #开盘价*100 stock_high = struct.unpack("l", stock_high) #最高价*100 stock_low= struct.unpack("l", stock_low) #最低价*100 stock_close = struct.unpack("l", stock_close) #收盘价*100 stock_amount = struct.unpack("f", stock_amount) #成交额 stock_vol = struct.unpack("l", stock_vol) #成交量 stock_reservation = struct.unpack("l", stock_reservation) #保留值 date_format = datetime.datetime.strptime(str(stock_date[0]),\'%Y%M%d\') #格式化日期 list= date_format.strftime(\'%Y-%M-%d\')+","+str(stock_open[0]/100)+","+str(stock_high[0]/100.0)+","+str(stock_low[0]/100.0)+","+str(stock_close[0]/100.0)+","+str(stock_vol[0])+"\r\n" file_object.writelines(list) file_object.close() stock_csv(\'D:/通达信/vipdoc/sh/lday/sh000001.day\', \'1\')
看了一下1.CSV 文件的输出数据，正确！
这么多年数据格式没变，也没加密。感谢通达信！祝通达信事业蒸蒸日上！
四、批量解析
import os import struct import datetime def stock_csv(filepath, name): data = [] with open(filepath, \'rb\') as f: file_object_path = \'D:/通达信/vipdoc/sh/pythondata/\' + name +\'.csv\' file_object = open(file_object_path, \'w+\') while True: stock_date = f.read(4) stock_open = f.read(4) stock_high = f.read(4) stock_low= f.read(4) stock_close = f.read(4) stock_amount = f.read(4) stock_vol = f.read(4) stock_reservation = f.read(4) # date,open,high,low,close,amount,vol,reservation if not stock_date: break stock_date = struct.unpack("l", stock_date) # 4字节 如20091229 stock_open = struct.unpack("l", stock_open) #开盘价*100 stock_high = struct.unpack("l", stock_high) #最高价*100 stock_low= struct.unpack("l", stock_low) #最低价*100 stock_close = struct.unpack("l", stock_close) #收盘价*100 stock_amount = struct.unpack("f", stock_amount) #成交额 stock_vol = struct.unpack("l", stock_vol) #成交量 stock_reservation = struct.unpack("l", stock_reservation) #保留值 date_format = datetime.datetime.strptime(str(stock_date[0]),\'%Y%M%d\') #格式化日期 list= date_format.strftime(\'%Y-%M-%d\')+","+str(stock_open[0]/100)+","+str(stock_high[0]/100.0)+","+str(stock_low[0]/100.0)+","+str(stock_close[0]/100.0)+","+str(stock_vol[0])+"\r\n" file_object.writelines(list) file_object.close() path = \'D:/通达信/vipdoc/sh/lday/\' listfile = os.listdir(\'D:/通达信/vipdoc/sh/lday/\') for i in listfile: stock_csv(path+i, i[:-4])
顺便写个简单的公式，其实与macd差不多
BASE:0; J10:SUM(C,10)/10 NODRAW; PL10:(C-J10)/J10*100 COLORSTICK; PL10SUM:IF(PL10>=0,SUM(PL10,BARSLASTCOUNT(PL10>=0)),SUM(PL10,BARSLASTCOUNT(PL10<0))); PL60:=(C-MA(C,60))/MA(C,60)*100; PL60SUM:IF(PL60>=0,SUM(PL60,BARSLASTCOUNT(PL60>=0)),SUM(PL60,BARSLASTCOUNT(PL60<0))) NODRAW; PL10DAY:IF(PL10>=0,BARSLASTCOUNT(PL10>=0),0-BARSLASTCOUNT(PL10<0)) NODRAW; MMAX:=MAX(MAX(MAX(MAX(MAX(MA(C,5),MA(C,10)),MA(C,20)),MA(C,30)),MA(C,60)),MA(C,120)); XL60:(MA(C,60)-REF(MA(C,60),10))/REF(MA(C,60),10)*100 NODRAW; PL250:(C-MA(C,250))/MA(C,250)*100 NODRAW; PL10SUMA:IF(C>MA(C,250),PL10SUM,1000) NODRAW;
参考：https://blog.csdn.net/weixin_33755554/article/details/86306394