1. Ken Pespisa04/08/2009 06:51:10 AM
Homepage: http://soprepare.blogspot.com

I've run up to this limitation in the past and the way I've solved it is to create an XML version of a view that outputs XML tags, most wrapped in CDATA blocks, and with Treat Contents as HTML turned on. Then I create a Page that wraps that view with the XML declaration and root element. In between the root element I put an embedded view. Finally I set the content-type to text\xml for the Page.

Then I can browse to that page and use the count=-1 parameter and get back any number of entries. I've had no problem (except a little sluggishness) getting back close to 10,000 records.

I realize this defeats the point of your attempts to use ReadViewEntries because this method requires some setup of the view. But just wanted to point out that there's another way if you run up on the documents limit obstacle. In many situations it would easier to just edit the view and create a page as I describe than to reconfigure code on the receiving end to take chunks of 1000 records at a time.

2. Gary04/08/2009 01:00:44 PM
Homepage: http://interoptips.com

Thanks Ken, I remember your method in a RedBook. It's a good approach. I have my Web Agent that work with parameters but folks tell me they are not allowed to put Agents on the server. You Page/View method maybe allowed.

3. Dwight Wilbanks04/22/2009 09:39:59 PM
Homepage: http://www.dwightwilbanks.org

I've worked long and hard to put lipsitck on that pig that is the domino view { Link } and have worked through quite a few of them, Attempting to make the web client as functional as the notes client. My first significant approach was using xslt, but, I quickly got ran out of options there. Of course there is the ?openagent approach that has had a zillion different makeovers.

For the most part, small views are easy any way that you choose, if you can push the whole thing down to the web client, you can do things like expand / collapse, sort to a different order. move, resize columns and of course have much more control over the actual look. the issue with that is that your still having to fully re-design the view in xslt or javascript (json) or whatever method you choose. not only are you having to redesign it, but you are redesinging in a fully proprietary environment, even all the other notes developers are not doing it the same way as you are.

What the open source project faceliftfordomino did was an attempt to make the web an accurate representation of the options that the developer selected in the designer client.

Think about this for a momemnt. With a DSAPI filter you could capture things like ?readviewentries and perform additional processing by doing a server side xslt, so the html would leave the server fully formatted with pretty html. The xslt could also be computer generated and so all you would need to do is issue the
?readviewenteries&format=xxx

After all that, when I go to full a client request for a view, after having performed and seen all these different approaches, I usually do it the old fashion way, pre Ajax , pre ReadViewEntries, pre json,