DOM guide: Working with documents

This section describes how to load COLLADA instance documents into the COLLADA runtime database and how to save the runtime database into COLLADA instance documents.

Note: In some cases, pointer and return value checks have been removed from this code for brevity.

Creating the DAE Object

To use the COLLADA DOM, your client code needs to include the following header files:

#include <dae.h>
#include <dom/domCOLLADA.h>

To use a specific effect profile, such as <profile_COMMON> or <profile_CG>, you must include the header for that profile. For example:

#include <dom/domProfile_CG.h>

A useful header to include is dom/domConstants.h. This provides constants for strings that are commonly used in the DOM. These strings include all the COLLADA element names and element type names. Many examples to follow use these constants.

After including the appropriate headers, your application needs to create a DAE object. The following example shows a DAE object created on the heap (using operator new) and accessed as a pointer.

DAE *collada_dom = new DAE();

Don't forget to call operator delete to delete the object when you're done. You could also create the DAE object globally or on the stack.

Loading Documents

Use the DAE::load method to read an existing COLLADA document into the runtime database. When you load a document, the COLLADA DOM uses the document URI as the name of the document when searching or doing other operations where the document needs to be identified. In this example, the name of the document created by the load process is file:///input_url_name.dae, the same as the URI it was loaded from:

daeInt error = collada_dom->load("file:///C:/Test/colladaDocument.dae");

Note: A common mistake is to pass a Windows or Linux file path to the DAE::load or DAE::save methods. See Using URIs in COLLADA.

The DAE::load method returns:

• DAE_OK if the load is successful.
• DAE_ERR_COLLECTION_ALREADY_EXISTS iff a document with the same name, or the same document, has already been loaded.
• DAE_ERR_BACKEND_IO for other errors encountered during load.

Note: The terms "collection" and "document" mean the same thing in the COLLADA DOM interface and documentation.

Creating New Documents

You can use the COLLADA DOM to create a set of elements, add them to an empty database, and write the database to a COLLADA instance document.

daeDocument *doc = NULL;
collada_dom->getDatabase()->insertDocument("file:///C:/Test/colladaDocument.dae", &doc);
// doc is now a valid COLLADA document

Saving Documents

After you have completed any changes to the elements in the document, you can write the document to a new URL using the DAE::save method. You can use either the document name or the document index to choose which collection to save:

daeInt error = collada_dom->save("file:///C:/Test/colladaDocument.dae");

To save the document to a different URI than the URI that was used to create it, use the DAE:saveAs method. The last argument to all of the save methods is an optional flag that specifies whether to overwrite an existing document. This flag defaults to true.

Note: The COLLADA DOM XML parser does not fully validate the data. It detects some types of errors in XML input but not all of them. You may want to run new COLLADA data through an XML validator before loading it into the DOM. If you are developing programs that write COLLADA data, you should run a validator on the output of your programs during testing. See schema validation.

Querying and Removing Documents

Three functions allow you to query for and iterate over documents in the runtime database:

• daeDatabase::getDocumentCount
• daeDatabase::getDocument
• daeDatabase::isDocumentLoaded

If you want to release a document from memory, you can use the DAE::unload method. The document to be unloaded is specified by its URI name. Removing a document removes the root element and, subsequently, the entire element hierarchy from the database. This method returns either:

• DAE_OK upon success.
• DAE_ERR_COLLECTION_DOES_NOT_EXIST if the document does not exist in the runtime database.

It's necessary to call DAE::unload only if you want the document released from memory immediately. Otherwise, the document is automatically removed when the parent DAE object is destroyed.

See also

• DOM load and save flow