Develop Reference

How Do I Deserialize Into Local Collections in .NET?

I am trying to build a data pipeline in .NET. I have been given an xsd and have used the XML Schema Definition Tool to generate C# classes that represent the object model. In order to load this data into my data store I need to transform the data coming out of the XML into structure that matches my application schema and collect/dedupe elements. To do this I have a parser class that will read a file and load the contents into local collections which will then be loaded into my database. I see two options to do this -
Manually loop through the XML with an XmlReader and pull out the data I need, loading it into the local collections in stream. This is not desirable because it does not take advantage of the strongly typed/strict xsd that I was given and requires a lot of hard coding things like while (reader.Read()), check for specific XML nodes, and then `reader.GetAttribute("HardCodedString").
Use XmlSerializer to deserialize the whole file at once and then loop through the deserialized collections and insert into my local collections. This is not desirable because the files could be very large and this method forces me to loop through all of the data twice (once to deserialize and once to extract the data to my local collections).
Ideally I would like some way to register a delegate to be executed as each object is deserialized to insert into my local collections. Is there something in the framework that allows me to do this? Requirements are as follows:
Performant - Only loop through the data once.
Functional - Data is inserted into the local collections during deserialization.
Maintainable - Utilize strongly typed classes that were generated via the xsd.
I have created a minimal example to illustrate my point.
Example XML File:
<Hierarchy xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://www.example.com/example">
<Children>
<Child ChildId="1" ChildName="First">
<Parents>
<Parent ParentId="1" ParentName="First" RelationshipStart="1900-01-01T00:00:00"/>
<Parent ParentId="2" ParentName="Second" RelationshipStart="2000-01-01T00:00:00"/>
</Parents>
</Child>
<Child ChildId="2" ChildName="Second">
<Parents>
<Parent ParentId="2" ParentName="Second" RelationshipStart="1900-01-01T00:00:00"/>
<Parent ParentId="3" ParentName="Third" RelationshipStart="2000-01-01T00:00:00"/>
</Parents>
</Child>
</Children>
</Hierarchy>
Local collections I am trying to load:
public Dictionary<int, string> Parents { get; }
public Dictionary<int, string> Children { get; }
public List<Relationship> Relationships { get; }
Manual version (not maintainable and doesn't use xsd):
public void ParseFileManually(string fileName)
{
using (var reader = XmlReader.Create(fileName))
{
while (reader.Read())
{
if (reader.NodeType == XmlNodeType.Element && reader.Name == "Hierarchy")
{
while (reader.Read())
{
if (reader.NodeType == XmlNodeType.Element && reader.Name == "Child")
{
int childId = int.Parse(reader.GetAttribute("ChildId"));
string childName = reader.GetAttribute("ChildName");
Children[childId] = childName;
while (reader.Read())
{
if (reader.NodeType == XmlNodeType.Element && reader.Name == "Parent")
{
int parentId = int.Parse(reader.GetAttribute("ParentId"));
string parentName = reader.GetAttribute("ParentName");
DateTime relationshipStart = DateTime.Parse(reader.GetAttribute("RelationshipStart"));
Parents[parentId] = parentName;
Relationships.Add(
new Relationship{
ParentId = parentId,
ChildId = childId,
Start = relationshipStart
});
}
else if (reader.NodeType == XmlNodeType.EndElement && reader.Name == "Child")
{
break;
}
}
}
}
}
}
}
}
Deserialize version (loops through the data twice):
public void ParseFileWithDeserialize(string fileName)
{
var serializer = new XmlSerializer(typeof(Hierarchy));
using (var fileStream = new FileStream(fileName, FileMode.Open))
{
var fileData = (Hierarchy) serializer.Deserialize(fileStream);
foreach (var child in fileData.Children)
{
Children[child.ChildId] = child.ChildName;
foreach (var parent in child.Parents)
{
Parents[parent.ParentId] = parent.ParentName;
Relationships.Add(
new Relationship
{
ParentId = parent.ParentId,
ChildId = child.ChildId,
Start = parent.RelationshipStart
});
}
}
}
}

You should use some annotations to get the data from the correct field in the XML, if you use these definitions;
public class Hierarchy
{
public Hierarchy()
{
Children = new List<Child>();
}
public List<Child> Children { get; set; }
}
public class Child
{
public Child()
{
Parents = new List<Parent>();
}
[XmlAttribute("ChildId")]
public int ChildId { get; set; }
[XmlAttribute("ChildName")]
public string ChildName { get; set; }
public List<Parent> Parents { get; set; }
}
public class Parent
{
[XmlAttribute("ParentId")]
public int ParentId { get; set; }
[XmlAttribute("ParentName")]
public string ParentName { get; set; }
[XmlAttribute("RelationshipStart")]
public DateTime RelationshipStart { get; set; }
}
Then you should be able to simplify your code to;
public static Hierarchy Deserialize(string fileName)
{
using (var fileStream = new StreamReader(fileName, Encoding.UTF8))
{
XmlSerializer ser = new XmlSerializer(typeof(Hierarchy));
return (Hierarchy)ser.Deserialize(fileStream);
}
}
To test it out you can create a sample data set and serialize it to a file, then use the above code to read it back
public static void Serialize(Hierarchy h, string fileName)
{
System.Xml.Serialization.XmlSerializer ser = new System.Xml.Serialization.XmlSerializer(typeof(Hierarchy));
StreamWriter sw = new StreamWriter(fileName, false, Encoding.UTF8);
ser.Serialize(sw, h);
}
Test Code
static void Test()
{
Hierarchy h = new Hierarchy();
Parent p1 = new Parent() { ParentId = 1, ParentName = "First", RelationshipStart = DateTime.Now };
Parent p2 = new Parent() { ParentId = 2, ParentName = "Second", RelationshipStart = DateTime.Now };
Parent p3 = new Parent() { ParentId = 3, ParentName = "Third", RelationshipStart = DateTime.Now };
Child c1 = new Child() { ChildId = 1, ChildName = "First" };
c1.Parents.Add(p1);
c1.Parents.Add(p2);
Child c2 = new Child() { ChildId = 2, ChildName = "Second" };
c2.Parents.Add(p2);
c2.Parents.Add(p3);
h.Children.Add(c1);
h.Children.Add(c2);
Serialize(h, AppContext.BaseDirectory + "Text.xml");
Hierarchy hReadBack = Deserialize(AppContext.BaseDirectory + "Text.xml");
}
Edit : To answer your question
Use these classes
public class Hierarchy
{
public Hierarchy()
{
Children = new List<Child>();
}
public List<Child> Children { get; set; }
private Dictionary<int, string> _parents;
private Dictionary<int, string> _childrenList;
private List<Relationship> _relationships;
private void CalcuateLists()
{
_parents = new Dictionary<int, string>();
_childrenList = new Dictionary<int, string>();
_relationships = new List<Relationship>();
foreach (Child c in this.Children)
{
if (!_childrenList.ContainsKey(c.ChildId))
{
_childrenList.Add(c.ChildId, c.ChildName);
}
foreach (Parent p in c.Parents)
{
if (!_parents.ContainsKey(p.ParentId))
{
_parents.Add(p.ParentId, p.ParentName);
}
if (_relationships.FirstOrDefault(dat => dat.ParentId == p.ParentId && dat.ChildId == c.ChildId) == null)
{
_relationships.Add(new Relationship() { ChildId = c.ChildId, ParentId = p.ParentId, Start = p.RelationshipStart });
}
}
}
}
public Dictionary<int, string> Parents {
get
{
if (_parents == null)
CalcuateLists();
return _parents;
}
}
public Dictionary<int, string> ChildrenList {
get
{
if (_childrenList == null)
CalcuateLists();
return _childrenList;
}
}
public List<Relationship> Relationships {
get
{
if (_relationships == null)
CalcuateLists();
return _relationships;
}
}
}
public class Child
{
public Child()
{
Parents = new List<Parent>();
}
[XmlAttribute("ChildId")]
public int ChildId { get; set; }
[XmlAttribute("ChildName")]
public string ChildName { get; set; }
public List<Parent> Parents { get; set; }
}
public class Parent
{
[XmlAttribute("ParentId")]
public int ParentId { get; set; }
[XmlAttribute("ParentName")]
public string ParentName { get; set; }
[XmlAttribute("RelationshipStart")]
public DateTime RelationshipStart { get; set; }
}
Then your test code becomes
public static void Test()
{
Hierarchy h = new Hierarchy();
Parent p1 = new Parent() { ParentId = 1, ParentName = "First", RelationshipStart = DateTime.Now };
Parent p2 = new Parent() { ParentId = 2, ParentName = "Second", RelationshipStart = DateTime.Now };
Parent p3 = new Parent() { ParentId = 3, ParentName = "Third", RelationshipStart = DateTime.Now };
Child c1 = new Child() { ChildId = 1, ChildName = "First" };
c1.Parents.Add(p1);
c1.Parents.Add(p2);
Child c2 = new Child() { ChildId = 2, ChildName = "Second" };
c2.Parents.Add(p2);
c2.Parents.Add(p3);
h.Children.Add(c1);
h.Children.Add(c2);
Serialize(h, AppContext.BaseDirectory + "Text.xml");
Hierarchy hReadBack = Deserialize(AppContext.BaseDirectory + "Text.xml");
Dictionary<int, string> Parents = hReadBack.Parents;
Dictionary<int, string> Children = hReadBack.ChildrenList;
List<Relationship> Relationships = hReadBack.Relationships;
}
EDIT
To get the results directly without looping
You will need this class
public class Relationship
{
public int ParentId { get; set; }
public int ChildId { get; set; }
public DateTime Start { get; set; }
}
And this selection
// Get a list of child ids and names
Dictionary<int, string> Children = (from c in hReadBack.Children select new { ChildId = c.ChildId, Name = c.ChildName}).ToDictionary(dat => dat.ChildId, dat => dat.Name);
// Get a parent ids and names
Dictionary<int, string> Parents = (from p in hReadBack.Children.SelectMany(i => i.Parents) select new { ParentId = p.ParentId, Name = p.ParentName }).Distinct().ToDictionary(dat => dat.ParentId, dat => dat.Name);
// Get the relationships
List<Relationship> Relationship = (from Child c in hReadBack.Children from Parent p in c.Parents select new Relationship() { ChildId = c.ChildId, ParentId = p.ParentId, Start = p.RelationshipStart }).ToList();

Search a hierarchy of class and return the path to get there

I have this class, a hierarchy of categories.
class Categories
{
public long Id { get; set; }
public long ParentId { get; set; }
public string Name { get; set; }
public bool IsActive { get; set; }
public List<Categories> ChildrenData { get; set; }
}
How can I recursively iterate through this class of unknown depth and return the path to get there?
All "Id" values are unique. Say I want to find Id = 23 and get the path to get there by concatenating "Name".
For example, in the image below searching for ID = 23 would return: Default Category/Books/Nonfiction/Best-sellers
Example Hierarchy

My suggestion is that you first build an index:
public static Dictionary<long, Category> IndexBuilder(Category c)
{
var index = new Dictionary<long, Category>();
IndexBuilder(c, index);
return index;
}
private static void IndexBuilder(Category c, Dictionary<long, Category> index)
{
if (index.ContainsKey(c.Id))
return;
index[c.Id] = c;
foreach(var child in c.ChildrenData)
IndexBuilder(child, index);
}
Now you have a lookup, and your path is then easy to produce:
static IEnumerable<Category> PathToRoot(long id, Dictionary<long, Category> index)
{
// Presumably the parent id of the top category is a sentinel.
long current = id
while (current != 0)
{
var category = index[current];
yield return category;
current = category.ParentId;
}
}
Or maybe we just go until we run out of index:
static IEnumerable<Category> PathToRoot(long id, Dictionary<long, Category> index)
{
long current = id
while (index.ContainsKey(current))
{
var category = index[current];
yield return category;
current = category.ParentId;
}
}
Now you have a tool you can use to make your string:
static string Slash<T>(this IEnumerable<T> items) =>
string.Join("/", items);
var s = PathToRoot(23, index)
.Reverse()
.Select(c => c.Name)
.Slash();
See what I am doing here? Make a bunch of helper methods each of which is about five lines long, that can be composed together to make powerful solutions.

I have provided 2 ways, the first way is recursive and the last is not.
Recursive way, add a reference to your parent. This way when you find a match you can easily traverse your way back up the chain to create your path.
class Categories
{
public Categories Parent { get; set; }
public long Id { get; set; }
public long ParentId { get; set; }
public string Name { get; set; }
public bool IsActive { get; set; }
public List<Categories> ChildrenData { get; set; }
}
Then add a Find() method:
public string Find(long id)
{
if( Id == id )
{
return GetPath(); //<-- we need to code this next.
}
else
{
foreach( var entry in Categories)
{
string path = entry.Find(id);
if( path != null )
{
return path;
}
}
return null;
}
}
And finally the GetPath(), the assumption here is that the highest level instances of Categories do not have a Parent:
public string GetPath()
{
System.Text.StringBuilder sb = new StringBuilder();
Categories current = this;
while( current != null)
{
sb.Insert(0,current.Name);
if( current != this)
{
sb.Insert(0,"/");
}
current = Parent;
}
return sb.ToString();
}
Now if recursion isn't what you want, then pass in the current path to the Find() method.
public string Find(long id, string pathSoFar)
{
if (pathSoFar == null)
{
pathSoFar = Name;
}
else
{
pathSoFar = pathSoFar + Name;
}
if ( Id == id)
{
return pathSoFar;
}
else
{
foreach( var entry in Categories)
{
string path = entry.Find(id, pathSoFar + "/");
if( path != null )
{
return path;
}
}
return null;
}
}
Usage:
var nonRecusive = cats.Find(23, null);

This will get what you are looking for using recursion:
void Main()
{
var data = GetData();
Console.WriteLine(GetPath(data, 23, ""));
}
public String GetPath(Categories c, Int32 id, String path)
{
if (c.Id == id)
{
return path + "/" + c.Name;
}
foreach (var cd in c.ChildrenData)
{
var p = GetPath(cd, id, path + "/" + c.Name);
if (!String.IsNullOrWhiteSpace(p))
{
return p;
}
}
return "";
}
public class Categories
{
public long Id { get; set; }
public long ParentId { get; set; }
public string Name { get; set; }
public bool IsActive { get; set; }
public List<Categories> ChildrenData { get; set; }
}
public Categories GetData()
{
return
new Categories
{
Id = 1,
Name = "Default Category",
ChildrenData = new List<Categories>
{
new Categories
{
Id = 2,
Name = "Magazines",
ChildrenData = new List<Categories> {}
},
new Categories
{
Id = 2,
Name = "Books",
ChildrenData = new List<Categories>
{
new Categories
{
Id = 20,
Name = "Fiction",
ChildrenData = new List<Categories> {}
},
new Categories
{
Id = 21,
Name = "Nonfiction",
ChildrenData = new List<Categories>
{
new Categories
{
Id = 22,
Name = "New",
ChildrenData = new List<Categories> {}
},
new Categories
{
Id = 23,
Name = "Best-Sellers",
ChildrenData = new List<Categories> {}
},
}
}
}
}
}
};
}

How would I search through all nested children of a custom class?

I have a table in a database which looks like this:
| id | parentID | name |
|----------+----------+-------------|
|ABCD-12345| | Top |
|----------+----------+-------------|
|ABCD-23456|ABCD-12345| Middle |
|----------+----------+-------------|
|ABCD-34567|ABCD-23456| Bottom |
|----------+----------+-------------|
|ABCD-45678|ABCD-23456| Bottom |
etc. - Basically, a hierarchical structure of N depth. I've taken this and shoved it into a datatable.
I have the following class built to hold this data:
public class TreeNode
{
public string id { get; set; }
public string name { get; set; }
public string parentID { get; set; }
public List<TreeNode> children { get; set; }
}
My goal is to go through each of these DataTable rows and insert them into the appropriate location in the TreeNode structure, but I'm super confused as to how I should approach this.
The main point of confusion for me is how I search through the entire existing structure of TreeNodes to see if a node with the parentID exists. Can anyone point me in the right direction?
I tried the following code, but it doesn't work:
public List<TreeNode> BuildTree(int currNode, List<TreeNode> treeList, DataTable dt)
{
foreach(DataRow row in dt.Rows)
{
if(row[1].ToString() == treeList[currNode].id)
{
treeList[currNode].children.Add(new TreeNode
{
id = row[0].ToString(),
name = row[2].ToString(),
parentID = row[1].ToString()
});
dt.Rows.Remove(row);
if(dt.Rows.Count > 0)
{
currNode++;
BuildTree(currNode, treeList, dt);
}
else
{
return treeList;
}
}
}
return null;
}
The problem is this line:
if(row[1].ToString() == treeList[currNode].id)
which gets an out of range exception, because I have a root at index 0, so on the second run (when currNode is 1), it breaks. I need to traverse to treeList[0].Children[int], followed by treeList[0].Children[int].Children[int] and so on and so forth.
So how do I accomplish this goal?

First I'm going to modify the TreeNode class for our convenience. It's not necessary, but just a nice to have. Also I'm going to assume that in your datatable you've done your error checking and there's only one node with ParentId = "".
public class TreeNode
{
public string Id { get; set; }
public string Name { get; set; }
public string ParentID { get; set; }
public List<TreeNode> Children { get; set; }
public TreeNode()
{
Id = Name = ParentID = string.Empty;
Children = new List<TreeNode>();
}
public bool IsRoot { get { return ParentID == string.Empty; } }
public bool IsChild { get { return Children == null || Children.Count == 0; } }
}
First, I'd convert your datatable data into a list of TreeNode objects. Forget about relationships, just create a list with each objects Children being empty. I wrote a method to simulate data retrival from datatable. Instead of that you can use your actual datatable.
static List<DataTableData> GetDataTableData()
{
var data = new List<DataTableData>
{
new DataTableData() { Id = "23456", ParentID = "12345", Name = "Middle" },
new DataTableData() { Id = "55555", ParentID = "12345", Name = "Middle" },
new DataTableData() { Id = "34567", ParentID = "23456", Name = "Bottom" },
new DataTableData() { Id = "12345", ParentID = string.Empty, Name = "Top" },
new DataTableData() { Id = "45678", ParentID = "23456", Name = "Bottom" },
new DataTableData() { Id = "66666", ParentID = "55555", Name = "Bottom" }
};
return data;
}
And this is what your Main() would look like:
static void Main(string[] args)
{
var treeNodes = new List<TreeNode>();
var dataTable = GetDataTableData();
foreach (var data in dataTable)
{
treeNodes.Add(new TreeNode() { Id = data.Id, Name = data.Name, ParentID = data.ParentID });
}
var root = BuildTree(treeNodes);
Console.ReadLine();
}
Now, in my BuildTree() method, instead of passing the datatable I can pass my TreeNode list, and return just the root node.
public static TreeNode BuildTree(List<TreeNode> nodes)
{
foreach (var node in nodes)
{
node.Children = nodes.Where(x => x.ParentID == node.Id).ToList();
}
return nodes.Find(x => x.IsRoot);
}
BuildTree() Breakdown
The nodes list already have all the nodes corresponding to data in your datatable. The BuildTree() is merely going to create the parent-child relations and fill in each object's Children list.
So I iterate through the list, and see what other elements in the list are supposed to be its children. When you have iterated through the list you'd created all the parent-child relationships. Finally, I pick the root node (the one who's ParentId is empty) and return it.
Edit
Here's an easy method to print and verify your tree.
static void PrintTree(TreeNode node, int indents)
{
for (int tab = 0; tab < indents; tab++)
{
Console.Write("\t");
}
Console.WriteLine("{0} - {1}", node.Id, node.Name);
if (node.Children != null && node.Children.Count > 0)
{
indents++;
foreach (var child in node.Children)
{
PrintTree(child, indents);
}
}
}
My output looks like this:

If you are building a class structure then you need a class with a recursive method. Not sure how efficient this will be if it gets too big. Execute the method from the top of the tree.
public class TreeNode
{
public string id { get; set; }
public string name { get; set; }
public string parentID { get; set; }
public List<TreeNode> children { get; set; }
public TreeNode() {
children = new List<TreeNode>();
}
public TreeNode FindParentWithID(string ID)
{
TreeNode ParentWithID = null;
//check my parentID if i am the one being looked for then return
if (id == ID) return this;
//search children
foreach (TreeNode treeNode in children)
{
ParentWithID = treeNode.FindParentWithID(ID);
if (ParentWithID != null)
{
break;
}
}
return ParentWithID;
}
}
You would load your data into the classes from the database. I had to hard code the values for the example to work:
TreeNode treeNode5 = new TreeNode() { id = "ABCD-12345", parentID = null, name = "Top" };
TreeNode treeNode6 = new TreeNode() { id = "ABCD-12346", parentID = "ABCD-12345", name = "Middle" };
treeNode5.children.Add(treeNode6);
TreeNode treeNode7 = new TreeNode() { id = "ABCD-12347", parentID = "ABCD-12346", name = "Bottom" };
TreeNode treeNode8 = new TreeNode() { id = "ABCD-12348", parentID = "ABCD-12346", name = "Bottom" };
treeNode6.children.Add(treeNode7);
treeNode6.children.Add(treeNode8);
TreeNode topOne = treeNode5.FindParentWithID("ABCD-12346");
topOne will be end up being treeNode6 name="Middle" in this example.

try this code, i have same issue and it works perfectly
this method used to build tree from list of items, by looping through all items, and add each item to its parent's child list. and return only the root item with its nested child.
public TreeNode BuildTree(List<TreeNode> source)
{
// build the children list for each item
foreach (var item in source)
{
var itm = source.Where(i => i.parentID == item.Id).ToList();
item.ChildItems = itm;
}
// return only the root parents with its child inside
return source.Where(i => i.parentID == null).FirstOrDefault();
}
noting that this method return only on TreeNode Object with its child, you can return List by changing .FirstOrDefault() to .ToList() in return line

Get the Depth of an object tree of objects with the same type using LAMBDA expression

I have this object:
public class dtHeader
{
public dtHeader ParentHeader { get; set; }
public string HeaderText { get; set; }
public string DataField { get; set; }
public bool Visible { get; set; }
public int DisplayOrder { get; set; }
}
I want to calculate using a lambda expression, the depth of the object, how many layers of the object in itself exists?
I saw this JavaScript post, but I am struggling to translate it to a one line lambda statement.
Lets say the object is as this new dtHeader(){ ParentHeader = null, HeaderText = "col1" };
the result would be 1
and for new dtHeader(){ ParentHeader = new dtHeader(){ ParentHeader = null, HeaderText = "col1" }, HeaderText = "col1" }; the result would be 2
I want to achieve this with a list<dtHeader>, so some of them would have a depth of 1 and others with deeper depths, and want the deepest depth.
_______ITEM_IN_LIST_OBJECT__
______1___2___3___4___5___6_
D 1. |_o_|_o_|_o_|_o_|_o_|_o_|
E 2. |_o_|___|_o_|___|_o_|_o_|
P 3. |___|___|_o_|___|_o_|___|
T 4. |___|___|___|___|_o_|___|
H 5. |___|___|___|___|_o_|___|
It must go infinitly(Until where it allows for objects to heap up inside eachother) deep.
var HeaderLayerCount = lDtCol.Where(n => n.ParentHeader != null)
.Where(n => n.ParentHeader.ParentHeader != null)
.Where(n => n.ParentHeader.ParentHeader.ParentHeader != null);
EDIT:
I just want to add that if you want to work on a specific depth level, for instance, all objects on a depth of 3, you can use this extra recursion function in the class
public class dtCol
{
public dtCol ParentHeader { get; set; }
public string HeaderText { get; set; }
public string DataField { get; set; }
public bool Visible { get; set; }
public int DisplayOrder { get; set; }
public int Depth { get { return ParentHeader != null ? ParentHeader.Depth + 1 : 1; } }
public int CurrentDepth { get; set; } //Set on initialisation
public dtCol getParent(dtCol col, int getDepth) //Gets the parent on a specific level after the first base level (1) else returns the previous not null child
{
return (col.ParentHeader != null && col.ParentHeader.CurrentDepth == getDepth) ? col.ParentHeader : this.getParent(col.ParentHeader, getDepth);
}
}
You can use it like so:
var HeaderLayerCount = lDtCol.OrderByDescending(n => n.Depth).First().Depth;
for (int hlc = 1; hlc <= HeaderLayerCount; hlc++)
{
var headerrow = new List<dtCol>();
//This foreach adds the parent header if not null else adds the not null child
lDtCol.ForEach(n =>
{
var h = n.getParent(n, hlc); //Get Parent, null is returned if parent does not exists
headerrow.Add((h != null) ? h : n); //If parent is null, add base dtCol so that the headers can be merged upwards.
});
//Do what you need with your new single dimensional list of objects
}

Why not implementing a int GetDepth() method on your class, that will reach the top most ancestor, counting each level?
Your query would then be much simpler.
I was outrunned by Frode, kudos to him
I had the same implementation:
public int GetDepth()
{
if (ParentHeader == null)
{
return 1;
}
else return 1 + ParentHeader.GetDepth();
}

using System;
using System.Linq;
namespace ConsoleApplication3
{
public class dtHeader
{
public dtHeader ParentHeader { get; set; }
public string HeaderText { get; set; }
public string DataField { get; set; }
public bool Visible { get; set; }
public int DisplayOrder { get; set; }
public int Depth
{
get
{
// If header has parent, then this depth is parent.depth + 1
if (ParentHeader != null)
return ParentHeader.Depth+1;
else
return 1; // No parent, root is depth 1
}
}
}
class Program
{
static void Main(string[] args)
{
dtHeader[] headers = {
new dtHeader { HeaderText = "dt1" },
new dtHeader { HeaderText = "dt2" },
new dtHeader { HeaderText = "dt3" },
new dtHeader { HeaderText = "dt4" },
new dtHeader { HeaderText = "dt5" }
};
headers[1].ParentHeader = headers[0];
headers[2].ParentHeader = headers[1];
headers[3].ParentHeader = headers[2];
headers[4].ParentHeader = headers[3];
var deepest = headers.OrderByDescending(item=>item.Depth).First();
Console.WriteLine(deepest.Depth+ ", " + deepest.HeaderText);
var runner = deepest;
while (runner.ParentHeader != null)
runner = runner.ParentHeader;
Console.WriteLine("The deepest root header is:" + runner.HeaderText);
}
}
}

Here's a lambda expression to get what you want:
Func<dtHeader, int> getDepth = null;
getDepth = dth =>
{
var depth = 1;
if (dth.ParentHeader != null)
{
depth += getDepth(dth.ParentHeader);
}
return depth;
};
You have to define it in two parts (assigning null & assigning the body) to let recursion work.

I modified Enigmativity's answer to make it work correctly:
Func<dtHeader, int, int> getDepth = null;
getDepth = (dth, depth) =>
{
if (dth.ParentHeader != null)
{
depth = getDepth(dth.ParentHeader, ++depth);
}
return depth;
};
Call it like this:
int depth = getDepth(header, 0)

How can I construct a tree using this data structure

I have this tree structure, which a node may have multiple nodes.
public class Node
{
public Node()
{
ChildLocations = new HashSet<Node>();
}
public int Id { get; set; }
public string Name { get; set; }
public virtual int? ParentLocationId { get; set; }
public virtual ICollection<Node> ChildLocations { get; set; }
}
Now, I want to add a list of parent-child value into this structure. like:
{1,A} -> {2,B}
{1,A} -> {3,C}
{1,A} -> {4,D}
{3,C} -> {5,E}
{3,C} -> {6,F}
to construct a tree looks like this:
1A
/ | \
2B 3C 4D
/ \
5E 6F
finally, it returns the root reference.
I have come out this solution. But I have no confidence with recursive part. It this correct?
public class Tree
{
Node root;
public Node Root
{
get { return root; }
}
public void Add(int parentId, string parentName, int childId, string childName)
{
if (root == null)
{
root = new Node { Id = parentId, Name = parentName };
root.ChildLocations.Add(new Node { Id = childId, Name = childName });
}
else
{
Add(root, parentId, parentName, childId, childName);
}
}
private void Add(Node node, int parentId, string parentName, int childId, string childName)
{
if (node == null)
{
return;
}
if (node.Id == parentId)
{
node.ChildLocations.Add(new Node { Id = childId, Name = childName });
return;
}
foreach (var n in node.ChildLocations)
{
Add(n, parentId, parentName, childId, childName);
}
}
}

As per my comment to your question, this works to build the tree you require:
public Node BuildTree()
{
var _1A = new Node() { Id = 1, Name = "A", };
var _2B = new Node() { Id = 2, Name = "B", };
var _3C = new Node() { Id = 3, Name = "C", };
var _4D = new Node() { Id = 4, Name = "D", };
var _5E = new Node() { Id = 5, Name = "E", };
var _6F = new Node() { Id = 6, Name = "F", };
_1A.ChildLocations.Add(_2B);
_1A.ChildLocations.Add(_3C);
_1A.ChildLocations.Add(_4D);
_3C.ChildLocations.Add(_5E);
_3C.ChildLocations.Add(_6F);
return _1A;
}
But it's not very general purpose. Can you elaborate on your needs?