Large data sets and performance

[houog]

 What are some essentials that I do to configure the chart in order to handle thousands of data points? We routinely graph and analyze up to 10,000 data points (Oil and Gas application).

[RandyJ]

Hi,

Performance of charts vary widely based on several factors such as datasources used, data size, server capabilities, data population methods, visual features, chart size and so on. It is usually related to the complexity of the chart being displayed, for example a Curve chart is much slower than a Line chart, Circle markers are slower than Squared ones.

Please note that many of the product's default settings are tailored for improving the chart's look rather than focusing on server performance. In situations where Chart FX is heavily used, these default settings may need to be configured to improve performance. In general, if server performance is of real concern you should refrain from making extensive use of the features geared towards achieving a cosmetic effect on the chart, particularly features like Borders, Transparency, Gradients, Motifs which heavily use the system graphical API as a way to improve the way charts look.

Regards,

Randy Jimenez 

[houog]

What data population methods are recommended for large data sets?

[JuanC]

In general more time is spent generating the visuals of the chart than processing the data, so the answer would actually depend on the structure your app uses to hold/retrieve the data, if you have your data in arrays/CLR objects we would still recommend you use binding - i.e. ItemsSource - because it will make easier to style if necessary (labels, markers, etc.)

It is preferable to use Chart.ItemsSource than Series.ItemsSource for large data sets.

JuanC