This is my first try into iPhone development and I have some UITableViewControllers which will show the data returned from a webservice.
In my AppDelegate I have setup a timer which is called every few seconds and reloads the model. I keep the same object reference and just refresh it’s contents, so I have the same object on any UITableViewController that is visible at the moment.
When the data is refreshed on the AppDelegate, I call:
[[(UITableViewController *)[self.navigationController topViewController] tableView] reloadData];
The model is basically a collection of file objects. Each file object has some properties and flags.
This works quite well if the current UITableViewController is a simple table, with one section, that maps each cell to a sequential index on the Array of files.
But I have one of the UITableViewControllers that shows this file details and, depending on a flag on the file object, will show 2 or 3 sections. I’ve put this logic here:
- (NSInteger)numberOfSectionsInTableView:(UITableView *)tableView {
if ([[_file status] intValue] & DOWNLOADING) {
kIndexTransferSpeed = 1;
kIndexSettings = 2;
return 3;
} else {
kIndexSettings = 1;
kIndexTransferSpeed = -1;
return 2;
}
}
So on this controller, my -tableView:cellForRowAtIndexPath: is just a big if statement that will show a property from the _file object, depending on the indexPath.
The problem with this approach is that when this table is reloading on the device, I can actually see it being reloaded. The section titles flash for a moment and the cells take a bit to show up. Specially if I am scrolling the table while it’s reloaded.
I’ve read somewhere that I should keep myself from putting any logic on those UITableViewController methods. But what should I do?
I thought about having a specific method that will create a “data view model” with the data ready to be consumed by - tableView:cellForRowAtIndexPath: and would be called by the AppDelegate before calling it’s [tableView reloadData]. But I’d have to recreate this whole “view model” everytime it’s called, since it’s the same object and I don’t really know what has been changed on the data model.
reloadData has to handle lots of things, including sections coming and going, objects moving in relative order etc. Without knowledge of what is actually going on there is sometimes nothing it can do but basically teardown and rebuilt the table.
If you want to avoid that you can instead inform the controller of the specific changes that have occured. Basically instead of your view reloading all the data through the dataSource in your controller, the controller tells it specifically what to up date and how.
It is a lot more work than just calling reload, but it is more efficient, uses less power, and because the UI has better information it not only avoids the whole thing disappearing and reappearing, but can actually animate in the changes.
The methods you want to look at are:
More documentation is available here.