如何以人类可读的格式记录 QuickFix 消息

Question

我想以标记名、值等解析模式记录 QuickFix 消息

我找不到现有的功能。我正在使用 QuickFix.Net。

我正在考虑提供某种方法来遍历所有呈现的标签并使用数据字典解析它们。

Answer 1

有一个 Java 示例http://www.quickfixj.org/confluence/display/qfj/Using+Message+Metadata可以转换为 .NET 并且可能证明是有用的：

public class MetadataExample {
    public static void main(String[] args) throws Exception {
        DataDictionary dd = new DataDictionary("FIX44.xml");

        Message m = new Message("8=FIX.4.4\0019=247\00135=s\00134=5\001"
                + "49=sender\00152=20060319-09:08:20.881\001"
                + "56=target\00122=8\00140=2\00144=9\00148=ABC\00155=ABC\001"
                + "60=20060319-09:08:19\001548=184214\001549=2\001"
                + "550=0\001552=2\00154=1\001453=2\001448=8\001447=D\001"
                + "452=4\001448=AAA35777\001447=D\001452=3\00138=9\00154=2\001"
                + "453=2\001448=8\001447=D\001452=4\001448=aaa\001447=D\001"
                + "452=3\00138=9\00110=056\001", dd);

        MessagePrinter printer = new MessagePrinter();
        printer.print(System.out, dd, m);
    }
}

public class MessagePrinter {

    public void print(DataDictionary dd, Message message) throws FieldNotFound {
        String msgType = message.getHeader().getString(MsgType.FIELD);
        printFieldMap("", dd, msgType, message.getHeader());
        printFieldMap("", dd, msgType, message);
        printFieldMap("", dd, msgType, message.getTrailer());
    }

    private void printFieldMap(String prefix, DataDictionary dd, String msgType, FieldMap fieldMap)
            throws FieldNotFound {

        Iterator fieldIterator = fieldMap.iterator();
        while (fieldIterator.hasNext()) {
            Field field = (Field) fieldIterator.next();
            if (!isGroupCountField(dd, field)) {
                String value = fieldMap.getString(field.getTag());
                if (dd.hasFieldValue(field.getTag())) {
                    value = dd.getValueName(field.getTag(), fieldMap.getString(field.getTag())) + " (" + value + ")";
                }
                System.out.println(prefix + dd.getFieldName(field.getTag()) + ": " + value);
            }
        }

        Iterator groupsKeys = fieldMap.groupKeyIterator();
        while (groupsKeys.hasNext()) {
            int groupCountTag = ((Integer) groupsKeys.next()).intValue();
            System.out.println(prefix + dd.getFieldName(groupCountTag) + ": count = "
                    + fieldMap.getInt(groupCountTag));
            Group g = new Group(groupCountTag, 0);
            int i = 1;
            while (fieldMap.hasGroup(i, groupCountTag)) {
                if (i > 1) {
                    System.out.println(prefix + "  ----");
                }
                fieldMap.getGroup(i, g);
                printFieldMap(prefix + "  ", dd, msgType, g);
                i++;
            }
        }
    }

    private boolean isGroupCountField(DataDictionary dd, Field field) {
        return dd.getFieldTypeEnum(field.getTag()) == FieldType.NumInGroup;
    }

}

输出：

BeginString: FIX.4.4
BodyLength: 247
MsgSeqNum: 5
MsgType: NewOrderCross (s)
SenderCompID: sender
SendingTime: 20060319-09:08:20.881
TargetCompID: target
SecurityIDSource: EXCHANGE_SYMBOL (8)
OrdType: LIMIT (2)
Price: 9
SecurityID: ABC
Symbol: ABC
TransactTime: 20060319-09:08:19
CrossID: 184214
CrossType: CROSS_TRADE_WHICH_IS_EXECUTED_PARTIALLY_AND_THE_REST_IS_CANCELLED (2)
CrossPrioritization: NONE (0)
NoSides: count = 2
  OrderQty: 9
  Side: BUY (1)
  NoPartyIDs: count = 2
    PartyIDSource: PROPRIETARY_CUSTOM_CODE (D)
    PartyID: 8
    PartyRole: CLEARING_FIRM (4)
    ----
    PartyIDSource: PROPRIETARY_CUSTOM_CODE (D)
    PartyID: AAA35777
    PartyRole: CLIENT_ID (3)
  ----
  OrderQty: 9
  Side: SELL (2)
  NoPartyIDs: count = 2
    PartyIDSource: PROPRIETARY_CUSTOM_CODE (D)
    PartyID: 8
    PartyRole: CLEARING_FIRM (4)
    ----
    PartyIDSource: PROPRIETARY_CUSTOM_CODE (D)
    PartyID: aaa
    PartyRole: CLIENT_ID (3)
CheckSum: 056

Answer 2

Quickfix 中没有方法以人类可读的格式解析消息。另一种选择是，当您在 onMessage 中处理传入的 FIX 消息时，您无论如何都会解析它以读取该消息。您可以在此处列出文件或数据库的标记名和值。但这样做的任何操作都可能会减慢 Quickfix 引擎的速度，因为写入文件或数据库总是很慢。所以要小心！

另一种选择是将消息记录到数据库而不是文件中，然后在数据库中执行所有操作，但这意味着在引擎和数据库中解析消息两次。但你会更加灵活地决定想读什么、不读什么。