2

I am using XML Deserialization for very large XML documents (sometimes over 200MB). The most documents are going very well, but sometimes the Deserialize function throws errors.

For example:

  • When an integer is a string, you get the "Wrong format error"
  • When a specified type is not recognized, you get an error.
Input string was not in a correct format.
   at System.Number.StringToNumber(String str, NumberStyles options, NumberBuffer& number, NumberFormatInfo info, Boolean parseDecimal)
   at System.Number.ParseUInt32(String value, NumberStyles options, NumberFormatInfo numfmt)
   at Microsoft.Xml.Serialization.GeneratedAssembly.XmlSerializationReaderFeatureCollection.Read208_Item(Boolean isNullable, Boolean checkType)
   at Microsoft.Xml.Serialization.GeneratedAssembly.XmlSerializationReaderFeatureCollection.Read209_Item(Boolean isNullable, Boolean checkType)
   at Microsoft.Xml.Serialization.GeneratedAssembly.XmlSerializationReaderFeatureCollection.Read210_Item(Boolean isNullable, Boolean checkType)
   at Microsoft.Xml.Serialization.GeneratedAssembly.XmlSerializationReaderFeatureCollection.Read211_Item(Boolean isNullable, Boolean checkType)
   at Microsoft.Xml.Serialization.GeneratedAssembly.XmlSerializationReaderFeatureCollection.Read212_Item(Boolean isNullable, Boolean checkType)
   at Microsoft.Xml.Serialization.GeneratedAssembly.XmlSerializationReaderFeatureCollection.Read213_Item(Boolean isNullable, Boolean checkType)
   at Microsoft.Xml.Serialization.GeneratedAssembly.XmlSerializationReaderFeatureCollection.Read228_TrailInformation(Boolean isNullable, Boolean checkType)
   at Microsoft.Xml.Serialization.GeneratedAssembly.XmlSerializationReaderFeatureCollection.Read237_FeatureCollectionFeatureMember(Boolean isNullable, Boolean checkType)
   at Microsoft.Xml.Serialization.GeneratedAssembly.XmlSerializationReaderFeatureCollection.Read238_FeatureCollection(Boolean isNullable, Boolean checkType)
   at Microsoft.Xml.Serialization.GeneratedAssembly.XmlSerializationReaderFeatureCollection.Read239_FeatureCollection()

The errors are no problem (I know how to handle them), it's very likely we will face other errors in the future which we cannot forsee now.

How can I tell the XMLSerializer to just continue and skip these errors?

I tried the custom eventHandlers, but they won't work.

I might be able to read the error messages and just delete the lines which are giving errors, but that's a really nasty way (See workaround below)!

Code:

try
{
    XmlSerializer xmlSerializer = new XmlSerializer(typeof(FeatureCollection));
    xmlSerializer.UnknownElement += new XmlElementEventHandler(Serializer_ElementEvent);
    xmlSerializer.UnknownAttribute += new XmlAttributeEventHandler(Serializer_AttributeEvent);
    xmlSerializer.UnknownNode += new XmlNodeEventHandler(Serializer_NodeEvent);
    xmlSerializer.UnreferencedObject += new UnreferencedObjectEventHandler(Serializer_UnreferencedObjectEvent);

    StringReader reader = new StringReader(xmlstring);
    Message = (FeatureCollection)xmlSerializer.Deserialize(reader); // This is where the exception takes place
}
catch (Exception e)
{
    System.Windows.MessageBox.Show("Something went wrong.", "Error", System.Windows.MessageBoxButton.OK);
    return;
}

Workaround (Work in Progress):

int tries = 1;
while (true)
{
    try
    {
        Bericht = (FeatureCollection)xmlSerializer.Deserialize(reader);
    }
    catch(Exception e)
    {
        if(tries >= 100)
        {
            throw e;
        }
        Exception innerException = e.InnerException;
        
        // REGEX linenumber innerException.Message; -> Error in XMLDocument (221, 6)
        // REMOVE LINE from source

        tries++;
        continue;
    }
    break;
}
Tomprg
  • 21
  • 2
  • *"The errors are no problem"* - how can error be not a problem? If you ignore *all errors*, then `Message` content will vary from 100% to 0% of what is inside xml. If you don't need `Message` at all (I guess), then why are you deserializing xml? – Sinatr Nov 20 '20 at 12:44
  • In this case i will introduce some kind of file validator. in next step process only documents without error and mark documents could not be processed. – MKasprzyk Nov 20 '20 at 13:27
  • The errors can be eliminated by commenting out the properties in the classes in c# which I usually do to isolate issues. You can fix conversion errors in custom get/set method. – jdweng Nov 20 '20 at 13:29
  • 2
    The errors are no problem in this context. I know how to solve them, I just don't know which errors I am going to face. I also know how to skip error giving files. My problem is, 99.99% of the XML file is just fine, it's only a couple of lines on 1.5 million lines XML. I want only to skip the 0.01% and keep the rest. – Tomprg Nov 20 '20 at 14:33

0 Answers0