Please advise the way forward

Question

My requirements are quite straightforward.
1. I want to store XML documents in the database as unstructured/LOB. I want to store other information about the documents so I am thinking of having a traditional relational data model, that contains an XMLType column all created manually.
2. I want to be able to browse the documents in the database (contained in the XMLType column of the manually created table) using HTTP (via web browser)
3. I want to be able to access the documents programmatically using HTTP
Questions
1. Given this scenario, do I need to register a schema? ( I think not)
2. How does one go about tying an XML DB Repository resource to the XMLType column in the relational data model?

Answer 1

if you want to use the XML DB HTTP server to acccess the content of the XML documents then
1. Fundamentally only content in an XMLType table or XMLType view can be associated with a resource. The reason for this is that a REF XMLType is used internally to track which document goes with which resource and a REF can only point at an object table.
Documents stored in an XMLType column in a relational table cannot (easily) be associated with a resource. You can get around this to some extent by having an XMLType view that exposes the each XMLType column in the relational table as an row in an XMLType view but this is messy.
2. If you want to insert or update via FTP / WebDAV you will need an XML Schema. The repository uses the SchemaLocation information to determine which underlying table should be updated when operations take place on the repository.
If the HTTP access is read only you will not need an XML Schema but you will have to use DBMS_XDB.CREATERESOUCE to create the resources in the repository. You will need to supply a REF XMLType that points to the correct row in the underlying table or view when creating the resource.
3. Have you considered using the Schema Based Metadata feature to the additional information about the documents you are managing. This basically allows the other information to be tracked as additional metadata associated with the resource. Basically you can define one (or more) XML Schemas that define the metadata you need to track and then you can associate content in the metadata tables with the documents in the repository. This has the nice side effect (from your perspective if you are still thinking of using SLIDE) that the metadata should be available via PROPFIND operatiions.
1. The documents will need to be accessable via the XML DB repository unless you want to restrict yourself to using the DBURI Servlet.