Tridion Core service - working with a Hierarchical Taxonomy - tridion

Question

I'm using the Tridion Core Service (Tridion 2011 SP1) to retrieve a list of keywords for given Category ID.
CoreService2010Client client = new CoreService2010Client();
XElement xmlCategoryKeywords = client.GetListXml(category.Id,
new KeywordsFilterData());
This returns what seems to be a flat XML structure representing our taxonomy which is 4 levels deep.
The documentation details an approach for working with this:
var categoryKeywords = xmlCategoryKeywords.Elements().Select(element =>
element.Attribute("ID").Value).Select(id => (KeywordData)client.Read(id, null)
);
foreach (KeywordData keyword in categoryKeywords)
{
Console.WriteLine("\t Keyword ID={0}, Title={1}", keyword.Id, keyword.Title);
}
However this will only list each Keyword. The KeywordData object contains property ParentKeywords so it would be possible to build the hierarchy in memory.
Is it possible to retrieve XML from the Core Service with a hierarchical structure? Or an easier way to work with this data?

Answer 1

One way is to use TaxonomiesOwlFilterData:
string publicationId = "tcm:0-3-1";
var filter = new TaxonomiesOwlFilterData();
filter.RootCategories = new[] {new LinkToCategoryData{ IdRef = "tcm:3-158-512"},};
var list = ClientAdmin.GetListXml(publicationId, filter);
As you see it is called on publication, but you can narrow it down to one or more categories. It will return you scary XML list that you can further process like this:
XNamespace tcmc = publicationId + "/Categories#";
XNamespace rdf = "http://www.w3.org/1999/02/22-rdf-syntax-ns#";
XNamespace tcmt = "http://www.tridion.com/ContentManager/5.2/Taxonomies#";
var taxonomyTree = new Dictionary<string, List<string>>();
var keywordNodes = list.Descendants(tcmc + "cat");
foreach (var keywordNode in keywordNodes)
{
var parents = new List<string>();
var parentNodes = keywordNode.Descendants(tcmt + "parentKeyword");
if (parentNodes.Count() > 0)
{
foreach (var parentNode in parentNodes)
{
parents.Add(parentNode.Attribute(rdf + "resource").Value);
}
}
taxonomyTree.Add(keywordNode.Attribute(rdf + "about").Value, parents);
}
As a result you will get unordered list of your keywords and corresponding parents that you can further process as you like. Item that has no parent is obviously a parent keyword. It might not be the most beatiful solution, but at least you will need only one call to server and not read each keyword.

Answer 2

You could process each branch, level by level. Here's some code I've been playing around with that does that:
CoreService2010Client client = new CoreService2010Client("basicHttp_2010");
KeywordsFilterData keywordsDataFilter = new KeywordsFilterData()
{
BaseColumns = ListBaseColumns.IdAndTitle,
IsRoot = true
};
UsingItemsFilterData usingItemsFilter = new UsingItemsFilterData()
{
BaseColumns = ListBaseColumns.IdAndTitle,
ItemTypes = new[] { ItemType.Keyword },
InRepository = new LinkToRepositoryData() { IdRef = "tcm:0-1-1" }
};
XElement parents = client.GetListXml("tcm:1-272-512", keywordsDataFilter);
foreach (XElement parent in parents.Descendants())
{
// Do something with the parent (top level) KW
XElement children = client.GetListXml(parent.Attribute("ID").Value, usingItemsFilter);
foreach (XElement child in children.Descendants())
{
// Do something with the child KW
}
}
I've found in the past that processing a flat list in to a hierarchy (in my case a list of all SGs in a Publication) created a massive overhead compared to processing a branch at a time. Of course I should caveat that by saying that I tried that with an old (early 5.x) version of Tridion so things may have improved since then.

Answer 3

Tridion 2011 SP1 comes with a new CoreService EndPoint. CoreService 2011. Its recommended to use the latest endpoint. Latest endpoint has new functionalists also bug fixes. SP1 also has a default coreservice client proxy that u can use directly in your code.