1. 深度行情接口OnRtnDepthMarketData的参数是这样的

struct CThostFtdcDepthMarketDataField
{
    ///交易日
    TThostFtdcDateType   TradingDay;
    ///合约代码
    TThostFtdcInstrumentIDType  InstrumentID;
    ///交易所代码
    TThostFtdcExchangeIDType ExchangeID;
    ///合约在交易所的代码
    TThostFtdcExchangeInstIDType    ExchangeInstID;
    ///最新价
    TThostFtdcPriceType  LastPrice;
    ///上次结算价
    TThostFtdcPriceType  PreSettlementPrice;
    ///昨收盘
    TThostFtdcPriceType  PreClosePrice;
    ///昨持仓量
    TThostFtdcLargeVolumeType   PreOpenInterest;
    ///今开盘
    TThostFtdcPriceType  OpenPrice;
    ///最高价
    TThostFtdcPriceType  HighestPrice;
    ///最低价
    TThostFtdcPriceType  LowestPrice;
    ///数量
    TThostFtdcVolumeType Volume;
    ///成交金额
    TThostFtdcMoneyType  Turnover;
    ///持仓量
    TThostFtdcLargeVolumeType   OpenInterest;
    ///今收盘
    TThostFtdcPriceType  ClosePrice;
    ///本次结算价
    TThostFtdcPriceType  SettlementPrice;
    ///涨停板价
    TThostFtdcPriceType  UpperLimitPrice;
    ///跌停板价
    TThostFtdcPriceType  LowerLimitPrice;
    ///昨虚实度
    TThostFtdcRatioType  PreDelta;
    ///今虚实度
    TThostFtdcRatioType  CurrDelta;
    ///最后修改时间
    TThostFtdcTimeType   UpdateTime;
    ///最后修改毫秒
    TThostFtdcMillisecType   UpdateMillisec;
    ///申买价一
    TThostFtdcPriceType  BidPrice1;
    ///申买量一
    TThostFtdcVolumeType BidVolume1;
    ///申卖价一
    TThostFtdcPriceType  AskPrice1;
    ///申卖量一
    TThostFtdcVolumeType AskVolume1;
    ///申买价二
    TThostFtdcPriceType  BidPrice2;
    ///申买量二
    TThostFtdcVolumeType BidVolume2;
    ///申卖价二
    TThostFtdcPriceType  AskPrice2;
    ///申卖量二
    TThostFtdcVolumeType AskVolume2;
    ///申买价三
    TThostFtdcPriceType  BidPrice3;
    ///申买量三
    TThostFtdcVolumeType BidVolume3;
    ///申卖价三
    TThostFtdcPriceType  AskPrice3;
    ///申卖量三
    TThostFtdcVolumeType AskVolume3;
    ///申买价四
    TThostFtdcPriceType  BidPrice4;
    ///申买量四
    TThostFtdcVolumeType BidVolume4;
    ///申卖价四
    TThostFtdcPriceType  AskPrice4;
    ///申卖量四
    TThostFtdcVolumeType AskVolume4;
    ///申买价五
    TThostFtdcPriceType  BidPrice5;
    ///申买量五
    TThostFtdcVolumeType BidVolume5;
    ///申卖价五
    TThostFtdcPriceType  AskPrice5;
    ///申卖量五
    TThostFtdcVolumeType AskVolume5;
    ///当日均价
    TThostFtdcPriceType  AveragePrice;
    ///业务日期
    TThostFtdcDateType   ActionDay;
};

其中 UpdateMillisec 为最后修改毫秒，int型

2. CtpMdApi中该接口函数中对日期处理错误、对毫秒处理不合适

错误和不合适之处已经改正，见注释：

    def onRtnDepthMarketData(self, data: dict) -> None:
        """行情数据推送"""
        # 过滤没有时间戳的异常行情数据
        if not data["UpdateTime"]:
            return

        # 过滤还没有收到合约数据前的行情推送
        symbol: str = data["InstrumentID"]
        contract: ContractData = symbol_contract_map.get(symbol, None)
        if not contract:
            return

        # 对大商所的交易日字段取本地日期
        if not data["ActionDay"] or contract.exchange == Exchange.DCE:
            # 这里废了那么大的劲，却使用了一个更新滞后的变量，属实不应该
            # self.current_date是由定时器几秒更新一次，
            # 对于一些跨夜品种，会导致几秒钟的tick的日期错误
            # date_str: str = self.current_date   
            date_str: str = datetime.now().strftime("%Y%m%d")    # hxxjava change
        else:
            date_str: str = data["ActionDay"]

        # 这里不好，为什么要故意降低接口的时间精度，放着毫秒不要而费劲地变化为0.1秒精度？
        # timestamp: str = f"{date_str} {data['UpdateTime']}.{int(data['UpdateMillisec']/100)}"        
        timestamp: str = f"{date_str} {data['UpdateTime']}." + str(data['UpdateMillisec']*1000).zfill(6) # hxxjava edit
        dt: datetime = datetime.strptime(timestamp, "%Y%m%d %H:%M:%S.%f")
        dt: datetime = CHINA_TZ.localize(dt)

        tick: TickData = TickData(
            symbol=symbol,
            exchange=contract.exchange,
            datetime=dt,
            name=contract.name,
            volume=data["Volume"],
            turnover=data["Turnover"],
            open_interest=data["OpenInterest"],
            last_price=adjust_price(data["LastPrice"]),
            limit_up=data["UpperLimitPrice"],
            limit_down=data["LowerLimitPrice"],
            open_price=adjust_price(data["OpenPrice"]),
            high_price=adjust_price(data["HighestPrice"]),
            low_price=adjust_price(data["LowestPrice"]),
            pre_close=adjust_price(data["PreClosePrice"]),
            bid_price_1=adjust_price(data["BidPrice1"]),
            ask_price_1=adjust_price(data["AskPrice1"]),
            bid_volume_1=data["BidVolume1"],
            ask_volume_1=data["AskVolume1"],
            gateway_name=self.gateway_name
        )

        if data["BidVolume2"] or data["AskVolume2"]:
            tick.bid_price_2 = adjust_price(data["BidPrice2"])
            tick.bid_price_3 = adjust_price(data["BidPrice3"])
            tick.bid_price_4 = adjust_price(data["BidPrice4"])
            tick.bid_price_5 = adjust_price(data["BidPrice5"])

            tick.ask_price_2 = adjust_price(data["AskPrice2"])
            tick.ask_price_3 = adjust_price(data["AskPrice3"])
            tick.ask_price_4 = adjust_price(data["AskPrice4"])
            tick.ask_price_5 = adjust_price(data["AskPrice5"])

            tick.bid_volume_2 = data["BidVolume2"]
            tick.bid_volume_3 = data["BidVolume3"]
            tick.bid_volume_4 = data["BidVolume4"]
            tick.bid_volume_5 = data["BidVolume5"]

            tick.ask_volume_2 = data["AskVolume2"]
            tick.ask_volume_3 = data["AskVolume3"]
            tick.ask_volume_4 = data["AskVolume4"]
            tick.ask_volume_5 = data["AskVolume5"]

        self.gateway.on_tick(tick)

1. date_str这里每个tick更新，会有额外性能开销的；
2. ctp的onRtnDepthMarketData函数推送一秒两次，修改前后tick.datetime不会有差别的。