I see a couple posts out here that look like they're having the same confusion I am, but I haven't seen a direct answer to this so here it goes...
1. The "virtual mode" talked about in the help is tied exclusively to the use of the WinDataSource object, correct? Looks that way in the help I'm browsing, just thought I'd verify.
2. There's no way to actually bind this object to a real database (i.e. an SQL table on a remote server for instance - real world client/server type stuff, not sample data generated in a two-d array or something).
3. Therefore there really is no virtual mode option available unless you're using only unbound grid functions with data you populate on your own.
Is this correct? That seems odd - other products I've seen (even on VS6) had the option of binding to a source object of sorts using the equivalent of a "firehose" cursor (forward only/read only) and the grid would show data just about immediately and continue loading information in the background - when opening extremely large tables this is handy - taking a massive "wait while I load 20,000 rows" hit is not an option. To me when I see that "million rows" example, this is what I'm assuming I have - but as near as I can tell that's not the case.
true?
thanks
-J
1. Yes, that's pretty much true. There may be other data sources out there that support loading data on demand, but as far as the Inbox and Infragistics controls go, the UltraDataSource is the only one I know of.
2. If you are using the UltraDataSource in OnDemand mode, then you don't bind it to anything. the component fires events to request data and you give it the data as it is requested. Where the data comes from is entirely up to you. So it could come from a SQL table or a remote server or anywhere else you like. It's entirely up to you.
3. The grid and the UltraDataSource can handle loading data on-demand. But as far as I know, The DotNet framework itself doesn't have any really great way to load data from the back end efficiently one row (or even several rows) at a time. You could, of course write queries to load a few rows at a time into memory. This would be useful if you had a very large set of data and didn't want to load it all at once.