Reading logic fails when encountered XML element with empty vaue

Question

I had this code for reading XML file:

using (XmlReader reader = XmlReader.Create(path + "\\AddressBook\\settings.xml"))
{
    while (reader.Read())
    {
        if (reader.IsStartElement())
        {
            if (reader.Name == "Person")
            {
                Person p = new Person();

                reader.Read();
                reader.Read();

                // Get name
                p.name = reader.Value;    
                reader.Read(); // skip end tag of previous element
                reader.Read(); // read start tag of email
                reader.Read(); // read email value

                // Get email
                p.email = reader.Value;    
                reader.Read();
                reader.Read();
                reader.Read();

                // Get first name
                p.address = reader.Value;
                reader.Read();
                reader.Read();
                reader.Read();

                // Get notes
                p.notes = reader.Value;    
                reader.Read();
                reader.Read();
                reader.Read();

                // Get dob
                p.dob = DateTime.Parse(reader.Value);

                lPeople.Add(p);
            }
        }
    }
}

Thing is it is working but failed on this XML element:

<Person>
    <Name /> 
    <Email /> 
    <StreetAddress /> 
    <Notes>note</Notes> 
    <DateOfBirth>5/20/2015 8:04:15 PM</DateOfBirth> 
</Person>

It failed because you can see when I added empty elements as name, email and address no start and end tags where added (by XMLWriter), rather just <Name/>, hence my reading logic above failed by reading wrong elements. What is the workaround about this usually?

note: Answer from here using extension method doesn't work for me for some reasons, still no start and end tags are written for empty elements

update: The answer by Jon Hanna below is not working for me. What happens is if I have read some element say "Email" as soon as ReadElementContentAsString is called, it seems it jumps on other element like "Notes". Btw this is link with the XML on which it fails J. Hanna's approach

Answer 1

A simple null check would do.

if(reader.value == null) 
    p.name = "";
else
    p.name = reader.value;

That being said this is not an ideal way of reading XML, you probably want to use LINQ to XML which allow for SQL query-like operations to retrieve data, or System.Xml.XmlDocument which is older than LINQ to XML but still in use as well. Here is a tutorial for that.

This SO Question also mentions how to use XMLDocuments to read an XML, you could use that approach as mentioned to retrieve data from XML to a person class, a null value is given if there is no value available. That way you do not have to hardcode your approach by manually doing reader.Read(), which your approach will also fail if the XML is slightly different with less/more elements.

This tutorial also goes over the basics of LINQ to XML with examples very similar to what you're trying to achieve.

EDIT after OP's comments

If this naive approach is the way you want to proceed, chances are to conform to some requirement like in an assignment, then just keep doing what you're doing but add a switch case to check what node you're dealing with.

if (reader.Value == null)
   continue;

switch(reader.Name)
{
   case "Email":
     person.Email = reader.Value;
     break;
   case "Address":
     person.Address = reader.Value;
     break;
}

Answer 2

It's not a matter of "workaround". You've written a parser that only accepts a given format, and have then passed in XML in another format.

For a start, if you can get empty elements, it would make sense to check reader.IsEmptyElement.

You'd still be expecting a certain number of text nodes between given elements though, which isn't very resilient.

I'd prefer something like:

using (var reader = XmlReader.Create(path + @"\AddressBook\settings.xml"))
  while (reader.Read())
    if (reader.IsStartElement() && reader.Name == "Person")
      people.Add(ReadPerson(reader.ReadSubtree()));

Where then ReadPerson() lets me worry about just that element at that point:

private static Person ReadPerson(XmlReader rdr)
{
  var person = new Person();
  using(rdr)
    while(rdr.Read())
      if(rdr.IsStartElement())
        switch(rdr.Name)
        {
          case "Name":
            person.Name = rdr.ReadElementContentAsString();
            break;
          case "Email":
            person.Email =  rdr.ReadElementContentAsString();
            break;
          case "StreetAddress":
            person.Address = rdr.ReadElementContentAsString();
            break;
          case "Notes":
            person.Notes = rdr.ReadElementContentAsString();
            break;
          case "DateOfBirth":
            person.DateOfBirth = DateTime.Parse(rdr.ReadElementContentAsString(), CultureInfo.GetCultureInfo("en-US"));
            break;
        }
  return person;
}

(Though if I knew who was responsible for this XML I would beg them to replace the date time from one intended for human beings raised in the US to one intended for XML parsing software: 2015-05-20T20:04:15.

This is resilient in the face of re-ordering of elements, and different amounts of whitespace between elements, as well as empty elements.

I'd also certainly consider replacing people.Add(ReadPerson(reader.ReadSubtree())) with yield ReadPerson(reader.ReadSubtree()). Then the calling code can create it's own list if it needs the list, but otherwise can use the Person objects as they come.