How to put for loop to a separate thread - iphone

Question

I have for loop that can read up to 100k rows.
When I start this function it block my UI until its done.
for (Item* sp in items){
data = [data stringByAppendingFormat:#"\"%#\",\"%#\","\n", ....];
}
How can I put this to separate thread so it doesn't block UI?

Answer 1

I don't think you've provided a complete example, but the simple use of Grand Central Dispatch should do the trick:
dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{
for (Item* sp in items){
data = [data stringByAppendingFormat:#"\"%#\",\"%#\","\n", ....];
}
// Then if you want to "display" the data (i.e. send it to any UI-element):
dispatch_async(dispatch_get_main_queue(), ^{
self.someControl.data = data;
});
// else simply send the data to a web service:
self.webService.data = data;
[self.webService doYourThing];
});

Answer 2

This may suits you
dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_BACKGROUND, 0), ^{
for(Item* sp in items)
{
data = [data stringByAppendingFormat:#"\"%#\",\"%#\","\n", ....];
}
dispatch_async(dispatch_get_main_queue(), ^{
//UI updates using item data
});
});

Answer 3

Best way to use NSInvocationOperation.
NSOperationQueue *OperationQueue=[[NSOperationQueue alloc] init];
NSInvocationOperation *SubOperation=[[NSInvocationOperation alloc] initWithTarget:self selector:#selector(LoopOperation) object:nil];
[SubOperation setQueuePriority:NSOperationQueuePriorityVeryHigh]; //You can also set the priority of that thread.
[OperationQueue addOperation:SubOperation];
-(void)LoopOperation
{
for (Item* sp in items)
{
data = [data stringByAppendingFormat:#"\"%#\",\"%#\","\n", ....];
}
dispatch_async(dispatch_get_main_queue(), ^{
//UI updates using item data
});
}

Answer 4

If order doesn't matter in your loop I prefer to use dispatch_apply() inside of dispatch_async(). The difference is that a traditional for(...) loop will put all of the work on a single thread, while dispatch_apply() will perform the individual iterations in parallel on multiple threads, but as a whole the loop is synchronous, where the loop will not exit until all of the processing is complete.
A good test to see if order matters in your loop is whether the loop can be performed backwards with the same results.