Content Services - List all descendants for parent node.

Hi,

We have a custom web UI that relies on Content Services security to determine which nodes a user has access to view. I'm trying to cache a collection of all descendant node IDs that a specific user can see starting from a parent node (root of the site).

I have found only 2 ways of getting out of the Content Services a list of all descendent nodes for a parent node.

1. Get a list of all node IDs for the site from the DTree table, then pass the entire list to the DocumentManagementService.GetNodes() method. This results in only one call to the Content Services, but with very large request and response data.

2. Traverse the list of nodes from DTree table and recursively call DocumentManagementService.ListNodes() on each branch. This results in a very large number of calls to Content Services.

Both operations take a very long time. Can anybody suggest a better way? Ideally I would pass only 1 node ID and get back a list of IDs for all descendants.

Much appreciated.

Regards,

Andrei

Find more posts tagged with

Comments

Kyle_Swidrovich

Have you considered extending the available web services with your own that implements a call specifically for what you need?

Andrei_Dounin

I didn't even know that was posssible - could you please point me to any samples/documentation on this?

I'm guessing there is currently no built-in way of achieving what we're looking for?

Regards,

Andrei

Kyle_Swidrovich

Hi Andrei,

Unlike LAPI, Content Web Services was designed to be extensible.
Adding this functionality would require the creation of an Oscript module to provide the server-side implementation and then automated scripts can be run that will generate the source necessary to generate a deployable web service exposing your new call(s).

There's a HelloWorld example that ships with core that shows you how one could accomplish this.

Information included in the documentation ("Programmer's Guide") is probably a good start.
Here's a link to the CS 10 version of the docs as hosted here in OTDN:

https://knowledge.opentext.com/knowledge/cs.dll?func=ll&objId=18047148&objAction=browse&viewType=1

Cheers,
Kyle

Carsten_Kulms

Not what you are asking for but: Would it be an option to cache only data for nodes respectively children of folders actually visited when these are visited? This would avoid cases where the application will load the information of thousands (or millions or more) of nodes in advance even when the user will never traverse that deep in the folder hierarchy.

Andrei_Dounin

I actually started with that initially - however making a web service call each time the user requested a new page to go deeper down the tree added a noticeable delay making the whole site appear slower. Also I ran into the issue of building a navigation menu that appears on every page that displays all the nodes down to level 4.

You just gave me an idea - maybe I will just cache the first 4 levels, and build the rest of the cache on demand. Unfortunately this will still mean slow navigation beyond level 4. Just to clarify, I'm caching only the node IDs (only the integer values), so even with a few million documents it shouldn't take up an unmanageable amount of memory.

Regards,

Andrei

Carsten_Kulms

I wonder if DocumentManagement.getNodeInContainer (available with CWS 10.0.0) would be useful: This "method returns the nodes under a container (including within subcontainers)" and takes a GetNodesInContainerOptions parameter which specifies MaxResults ("the maximum number of results to return") and MaxDepth ("the maximum number of containers to recursively search").