This question is locked. New answers and comments are not allowed.
Hello,
I am having some trouble with the performance of the RadJumpList control when adding items to a collection that has already been bound to the control. I have uploaded an example project that shows this problem here.
In this project, there are two buttons, Unbound and Bound. Both buttons create an ObservableCollection of 100 objects randomly assigned to one of three headers, but the Bound button binds the collection to the RadJumpList before adding the items.
On my device, the unbound execution is very fast, under 1 second, but the bound execution takes about 8 seconds. If you increase the number of items to 1000, the unbound execution still completes in under a second but the bound execution is much slower, taking over a minute to complete.
I noticed that removing the group descriptor results in much improved performance for the bound scenario. The group descriptor is currently using string comparison, but I noticed no significant difference in execution time when using an integer instead of a string.
For comparison, I have not had this issue with a similarly configured LongListSelector from the SL toolkit. The LongListSelector has a much different collection format though (a collection of collections rather than using a group descriptor), so that may be part of the reason.
Is there any way to improve the performance when adding items to a bound collection?
Thanks,
Matt
I am having some trouble with the performance of the RadJumpList control when adding items to a collection that has already been bound to the control. I have uploaded an example project that shows this problem here.
In this project, there are two buttons, Unbound and Bound. Both buttons create an ObservableCollection of 100 objects randomly assigned to one of three headers, but the Bound button binds the collection to the RadJumpList before adding the items.
On my device, the unbound execution is very fast, under 1 second, but the bound execution takes about 8 seconds. If you increase the number of items to 1000, the unbound execution still completes in under a second but the bound execution is much slower, taking over a minute to complete.
I noticed that removing the group descriptor results in much improved performance for the bound scenario. The group descriptor is currently using string comparison, but I noticed no significant difference in execution time when using an integer instead of a string.
For comparison, I have not had this issue with a similarly configured LongListSelector from the SL toolkit. The LongListSelector has a much different collection format though (a collection of collections rather than using a group descriptor), so that may be part of the reason.
Is there any way to improve the performance when adding items to a bound collection?
Thanks,
Matt