Last Updated: 02 Sep 2016 09:04 by ADMIN
Created on: 09 Aug 2016 07:14
Category: UI for WPF
Type: Feature Request
Ordered dataset for ultraefficient trackbals engine and pan/zom behavior
If series datasource is ordered, asc or desc, all actions that serch a subset of data can be ultra speed. At the moment working with large datasets and zoom/trackbals cause occasional chart freeze.

Possible solution:

1) Declarative: The developer declare that the dataset is ordered asc/desc.
2) Automatic: On points add/remove chart verifies if data is ordered.
3) Programming: Chart allow to override serch in database functions.

My suggestion is 1) or/and 3)


DECLINED: Duplicate with
Posted on: 02 Sep 2016 09:04
Thank you for the detailed information on this matter, Marcello. This item is marked as DECLINED due to duplication with the one you mentioned in your comment above - . You can track its progress and vote for it in order to raise its priority.
Posted on: 15 Aug 2016 09:27
I try to explain more accurately the request.

Suppose that we know that a series dataset is ordered respect X.
If is, every procedure that search points closer to an X point can be extremely efficient.
In example, this happens drawing trackbals or showing labels and perhaps determining the points subset for a X pan/zoom.

Now, chart can choose to use efficient search alghoritms only if dataset is ordered.
I purposed three possible solution about this:

1) Declarative. Developer declare that a Series Dataset is ordered using a new series property. In Example IsPointsXOrdered = [None (default), Asc, Desc]. I the property is set Asc or Desc chart can use efficient alghoritms.

2) Automatic: If the same property can assume the "Automatic" value, additing/moving points in datase start a subroutine that determine if dataset is ordered or not. This solution can be extremely expensive and I don't suggest it.

3) Programming. If search points functions is set public virtual, developer can be implement a own search logic assuming the dataset is ordered,partially ordered or not ordered.

I think the best solution is 1) + 3).
1) is perfect for a lot of ordered case (In example dataset is database side ordered).
3) is perfect for more complex scenario.

Point 3) is already approved at
Posted on: 10 Aug 2016 07:11
Approved solution 3 at