예제 #1
0
def invoke_daily(ts_code, start_date, end_date):
    invoke_daily = pro.daily(ts_code=ts_code,
                             start_date=start_date,
                             end_date=end_date)
    return invoke_daily
예제 #2
0
def import_tushare_stock_daily(chain_param=None, ts_code_set=None):
    """
    插入股票日线数据到最近一个工作日-1。
    如果超过 BASE_LINE_HOUR 时间,则获取当日的数据
    :return:
    """
    table_name = 'tushare_stock_daily_md'
    logging.info("更新 %s 开始", table_name)

    has_table = engine_md.has_table(table_name)
    # 进行表格判断,确定是否含有tushare_stock_daily
    if has_table:
        sql_str = """
            SELECT ts_code, date_frm, if(delist_date<end_date, delist_date, end_date) date_to
            FROM
            (
            SELECT info.ts_code, ifnull(trade_date, list_date) date_frm, delist_date,
            if(hour(now())<16, subdate(curdate(),1), curdate()) end_date
            FROM 
                tushare_stock_info info 
            LEFT OUTER JOIN
                (SELECT ts_code, adddate(max(trade_date),1) trade_date FROM {table_name} GROUP BY ts_code) daily
            ON info.ts_code = daily.ts_code
            ) tt
            WHERE date_frm <= if(delist_date<end_date, delist_date, end_date) 
            ORDER BY ts_code""".format(table_name=table_name)
    else:
        sql_str = """
            SELECT ts_code, date_frm, if(delist_date<end_date, delist_date, end_date) date_to
            FROM
              (
                SELECT info.ts_code, list_date date_frm, delist_date,
                if(hour(now())<16, subdate(curdate(),1), curdate()) end_date
                FROM tushare_stock_info info 
              ) tt
            WHERE date_frm <= if(delist_date<end_date, delist_date, end_date) 
            ORDER BY ts_code"""
        logger.warning('%s 不存在,仅使用 tushare_stock_info 表进行计算日期范围', table_name)

    with with_db_session(engine_md) as session:
        # 获取每只股票需要获取日线数据的日期区间
        table = session.execute(sql_str)
        # 计算每只股票需要获取日线数据的日期区间
        begin_time = None
        # 获取date_from,date_to,将date_from,date_to做为value值
        code_date_range_dic = {
            ts_code:
            (date_from if begin_time is None else min([date_from, begin_time]),
             date_to)
            for ts_code, date_from, date_to in table.fetchall()
            if ts_code_set is None or ts_code in ts_code_set
        }

    # data_len = len(code_date_range_dic)
    data_df_list, data_count, all_data_count, data_len = [], 0, 0, len(
        code_date_range_dic)
    logger.info('%d stocks will been import into tushare_stock_daily_md',
                data_len)
    # 将data_df数据,添加到data_df_list

    try:
        for num, (ts_code, (date_from,
                            date_to)) in enumerate(code_date_range_dic.items(),
                                                   start=1):
            logger.debug('%d/%d) %s [%s - %s]', num, data_len, ts_code,
                         date_from, date_to)
            df = pro.daily(ts_code=ts_code,
                           start_date=datetime_2_str(date_from,
                                                     STR_FORMAT_DATE_TS),
                           end_date=datetime_2_str(date_to,
                                                   STR_FORMAT_DATE_TS))
            data_df = df
            if len(data_df) > 0:
                while try_2_date(df['trade_date'].iloc[-1]) > date_from:
                    last_date_in_df_last, last_date_in_df_cur = try_2_date(
                        df['trade_date'].iloc[-1]), None
                    df2 = pro.daily(ts_code=ts_code,
                                    start_date=datetime_2_str(
                                        date_from, STR_FORMAT_DATE_TS),
                                    end_date=datetime_2_str(
                                        try_2_date(df['trade_date'].iloc[-1]) -
                                        timedelta(days=1), STR_FORMAT_DATE_TS))
                    if len(df2 > 0):
                        last_date_in_df_cur = try_2_date(
                            df2['trade_date'].iloc[-1])
                        if last_date_in_df_cur < last_date_in_df_last:
                            data_df = pd.concat([data_df, df2])
                            df = df2
                        elif last_date_in_df_cur == last_date_in_df_last:
                            break
                        if data_df is None:
                            logger.warning(
                                '%d/%d) %s has no data during %s %s', num,
                                data_len, ts_code, date_from, date_to)
                            continue
                        logger.info('%d/%d) %d data of %s between %s and %s',
                                    num, data_len, data_df.shape[0], ts_code,
                                    date_from, date_to)
                    else:
                        break
            #把数据攒起来
            if data_df is not None and data_df.shape[0] > 0:
                data_count += data_df.shape[0]
                data_df_list.append(data_df)

            # 大于阀值有开始插入
            if data_count >= 500:
                data_df_all = pd.concat(data_df_list)
                # tot_data_df.to_sql(table_name, engine_md, if_exists='append', index=False, dtype=dtype)
                bunch_insert_on_duplicate_update(data_df_all, table_name,
                                                 engine_md,
                                                 DTYPE_TUSHARE_STOCK_DAILY_MD)
                all_data_count += data_count
                data_df_list, data_count = [], 0

                # # 数据插入数据库
                # data_df_all = data_df
                # data_count = bunch_insert_on_duplicate_update(data_df_all, table_name, engine_md, DTYPE_TUSHARE_STOCK_DAILY_MD)
                # logging.info("更新 %s 结束 %d 条信息被更新", table_name, data_count)
                # data_df = []
            # 仅调试使用
            # n=1
            # n=n+1
            # if DEBUG and n > 10:
            #     break
    finally:
        # 导入数据库
        if len(data_df) > 0:
            data_df_all = data_df
            data_count = bunch_insert_on_duplicate_update(
                data_df_all, table_name, engine_md,
                DTYPE_TUSHARE_STOCK_DAILY_MD)
            logging.info("更新 %s 结束 %d 条信息被更新", table_name, all_data_count)
            if not has_table and engine_md.has_table(table_name):
                alter_table_2_myisam(engine_md, [table_name])
                build_primary_key([table_name])