What is the fastest/most efficient way to read this XML to Dictionary (Linq or something else?)

16,894

Solution 1

When are are using a event based interface, similar to the one presented in your update, you will need to remember the name of the previous start element event. Often it is worth while holding a stack to keep track of the events. I would probably do something similar to the following:

public class PriceLevel
{
    private decimal? bid = null;
    private decimal? offer = null;

    public decimal? Bid {
        get { return bid; }
        set { bid = value; }
    }

    public decimal? Offer {
        get { return offer; }
        set { offer = value; }
    }
}

public delegate void OnPriceChange(long instrumentId, Dictionary<decimal, PriceLevel> prices);

public class MainClass
{
    private Stack<String> xmlStack = new Stack<String>();
    private Dictionary<decimal, PriceLevel> prices = new Dictionary<decimal, PriceLevel>();
    private bool isBids = false;
    private decimal? currentPrice = null;
    private long instrumentId;
    private OnPriceChange _priceChangeCallback;

    public void MainClass(OnPriceChange priceChangeCallback) {
        this._priceChangeCallback = priceChangeCallback;
    }

    public void XmlStart(object source, MessageEventArgs args) {
        xmlStack.Push(args.Value);

        if (!isBids && "bids" == args.Value) {
            isBids = true;
        }
    }

    public void XmlEnd(object source, MessageEventArgs args) {
        xmlStack.Pop();

        if (isBids && "bids" == args.Value) {
            isBids = false;
        }

        // Finished parsing the orderBookEvent
        if ("orderBook" == args.Value) {
            _priceChangeCallback(instrumentId, prices);
        }
    }

    public void XmlContent(object source, MessageEventArgs args) {

        switch (xmlStack.Peek()) {
        case "instrumentId":
            instrumentId = long.Parse(args.Value);
            break;

        case "price":
            currentPrice = decimal.Parse(args.Value);
            break;

        case "quantity":

            if (currentPrice != null) {
                decimal quantity = decimal.Parse(args.Value);

                if (prices.ContainsKey(currentPrice)) {
                    prices[currentPrice] = new PriceLevel();
                }
                PriceLevel priceLevel = prices[currentPrice];

                if (isBids) {
                    priceLevel.Bid = quantity;
                } else {
                    priceLevel.Offer = quantity;
                }
            }
            break;
        }
    }
}

Solution 2

first you need to get all offers and all bids

XDocument xmlDoc = XDocument.Load("TestFile.xml");


var bids = (from b in xmlDoc.Descendants("bids")
           select b).ToList();

var offers = (from o in xmlDoc.Descendants("offers")
           select o).ToList();

then you just iterate throgh bids and offers and add them to the dictionary... but as someone sait before... you will maybe have the problem that an pricelevel will have both bids and offers set if they have the same price

to iterate throgugh the list you just do this

foreach (XElement e in bids)
{
   price = e.Element("price").Value;
   quantity = e.Element("quantity").Value;
   dictionary.add(price, new PriceLevel(quantity,null);
}

the same you do for offer... but again.. .you probably have to check if this key already exists...

Share:
16,894
Macin
Author by

Macin

Updated on June 23, 2022

Comments

  • Macin
    Macin almost 2 years

    I am very new to parsing XML, and I started learning about linq which I think might be the best solution here. I am mostly interested in performance as the application I am creating will read stock exchange prices, which sometimes can change very rapidly. I receive following message from the server:

    <?xml version="1.0" encoding="utf-16"?>
        <events>
            <header>
                <seq>0</seq>
            </header>
            <body>
                <orderBookStatus>
                    <id>100093</id>
                    <status>Opened</status>
                </orderBookStatus>
                <orderBook>
                    <instrumentId>100093</instrumentId>
                    <bids>
                        <pricePoint>
                            <price>1357.1</price>
                            <quantity>20</quantity>
                        </pricePoint>
                        <pricePoint>
                            <price>1357.0</price>
                            <quantity>20</quantity>
                        </pricePoint>
                        <pricePoint>
                            <price>1356.9</price>
                            <quantity>71</quantity>
                        </pricePoint>
                        <pricePoint>
                            <price>1356.8</price>
                            <quantity>20</quantity>
                        </pricePoint>
                    </bids>
                    <offers>
                        <pricePoint>
                            <price>1357.7</price>
                            <quantity>51</quantity>
                        </pricePoint>
                        <pricePoint>
                            <price>1357.9</price>
                            <quantity>20</quantity>
                        </pricePoint>
                        <pricePoint>
                            <price>1358.0</price>
                            <quantity>20</quantity>
                        </pricePoint>
                        <pricePoint>
                            <price>1358.1</price>
                            <quantity>20</quantity>
                        </pricePoint>
                        <pricePoint>
                            <price>1358.2</price>
                            <quantity>20</quantity>
                        </pricePoint>
                    </offers>
                    <lastMarketClosePrice>
                        <price>1356.8</price>
                        <timestamp>2011-05-03T20:00:00</timestamp>
                    </lastMarketClosePrice>
                    <dailyHighestTradedPrice />
                    <dailyLowestTradedPrice />
                    <valuationBidPrice>1357.1</valuationBidPrice>
                    <valuationAskPrice>1357.7</valuationAskPrice>
                    <lastTradedPrice>1328.1</lastTradedPrice>
                    <exchangeTimestamp>1304501070802</exchangeTimestamp>
                </orderBook>
            </body>
        </events>
    

    My aim is to parse price point elements

    <pricePoint>
          <price>1358.2</price>
          <quantity>20</quantity>
    </pricePoint>
    

    into dictionary of the following structure:

    Dictionary<double, PriceLevel> 
    

    where price should be a double and PriceLevel is a class

    class PriceLevel
    {
         int bid;
         int offer;
    
         public PriceLevel(int b, int o)
         {
              bid = b;
              offer = o;
         }
    
    
    }
    

    Depending on the element, in which each price point exists (either bids or offers) quantity should be assigned accordingly, i.e. if price point exists in bids, then quantity should be assigned to bid, and 0 to offer. On the opposite, if price point exists in offers, then quantity should be assigned to offer and 0 to bid.

    I hope my explanation is clear, however if you have any problems understanding it, please do not hesitate to ask for clarification in comments. I would greatly appreciate help in solving this problem.

    +++++++++++++++++++++++++++++++++++++++++ Update:

    I have gone deeper into the stream I am trying to read, and it is not going to be as simple as I expected. I found out, that the stream will not always contain the whole document, therefore I will have to read it using XmlReader to process the stream on the ongoing basis. In this case, how can I read bids and offers? I have something like this:

    StreamReader sr = new StreamReader("..\..\videos.xml");

            XmlReader xmlReader = XmlReader.Create(sr);
            while (xmlReader.Read())
            {
                if (xmlReader.HasValue)
                {
                    OnXmlValue(this, new MessageEventArgs(true, xmlReader.Value));//saxContentHandler.Content(xmlReader.Value);
                }
                else
                {
                    if (xmlReader.IsEmptyElement)
                    {
                        OnStartElement(this, new MessageEventArgs(false, xmlReader.Name));
                        OnEndElement(this, new MessageEventArgs(false, xmlReader.Name));
                    }
                    else if (xmlReader.IsStartElement())
                    {
                        OnStartElement(this, new MessageEventArgs(false, xmlReader.Name));
                    }
                    else
                    {
                        OnEndElement(this, new MessageEventArgs(false, xmlReader.Name));
                    }
                }
            }
    

    but I am struggling to link element name to its value ... ie, how can I know which bid price point I am currently reading and if this exists in bids or offers? Thank you for help

  • Matt Ellen
    Matt Ellen about 13 years
    +1 for key collision, -1 for suggesting XDocument, without knowing if it is most efficient.
  • Macin
    Macin about 13 years
    Thank you Michael, this is very useful answer!