What is the best way to explore a hierarchical folder structure?

Question

Hallo,
I need to access and navigate a hierarchical folder structure hosted in a MS SQL Server database. In particular there is a root folder containing several folders. Each child-folder contains further nested folders or documents.
For each item I need to retrieve the folder's (name, path, etc) and the documents (title, author, etc.) details that are retrievable from the DB fields. Afterwards I will use these data to create a semantic web ontology using Jena API.
My question was about which is the best way to proceed.
A collegue of mine suggested to use the "WITH" command of SQL Server to create and use a link list to navigate easily the structure, executing just one query rather than several (one for each level of the nested loops). However in this way the solution will work only with the SMQ Server database, while my goal is to achieve a more general solution.
May someone help me?
Thank you in advance,
Francesco

Answer 1

My goal is to create a documents library ontology achieving, from each element of the hierarchy (folder or document), some data (title, parent, etc.) and use them to "label" the ontology resources.
I will use a little of both approches in the following way:
1) I make just ONE query on folder table to get, from each folder, its path (eg. root/fold1/fold2/doc1.pdf), its ID and ParentID and ONE on the Documents table to get the containerID, title, etc.
2) I create as many Folder objects as the retrieved records and an HashTable, where the KEY = Folder.ParentID value and the VALUE = Vector<Folder>. I add then each object to the Vector relative to the same ParentID. In this way I have an Vector containing all the folders child of the same parent folder and I do the same for an HashTable keeping the documents contained in a specific folder.
3)I extract from the HashTable the root folder (whose ParentID is always "0" and whose ID is "1") than it is invoked the method appendChild() (see code)
     public static void appendChild(int ID, Resource RES)
          Vector<Folder> currFold = table.get(ID);
          for(int i=0; i<currFold.size(); i++)
               //Extract the child and crate the relative resource
               if(table.containsKey(currFold.getID()))
                    appendChild(currFold[i].getID(), Resource newRES);
In this way I go in depth in the hirarchical structure using a "left most" procedure. I made a test and the output is correct. However, such an approch must be done for about 4 level depth folders (around 30 in total) containing also documents and creating the documents library of a Project. Then I must process around 20 project to achieve for all of them such a documents library representation.
By the way, I do not have to mantein the HashTable content after I created the docs library ontology. Hence I use just one hashTable for ALL the projects and I flush it after I finish to do the loop for one project in order to save resources.
My question is: is right my approach or might I improve it in some way?
Thank you for every suggesion/comment.
Francesco
Edited by: paquito81 on May 27, 2008 8:15 AM