Background

When you launch the Analysis screen and select Actions in the Display dropdown, Pentana will draw a chart showing all audits, each bar representing a count of the number of actions.

Once the database contains a certain number of audits this can cause slowness or even an error, due to the number of bars required on screen.  Exactly where this tipping point will be cannot be predicted, as the performance relates to speed of hardware and current load.

Appropriate use of the Chart Point Limit setting, pre-filters and views will help mitigate the issue.

Example

In a dataset where the number of data points is large you would normally first see a dialog box stating a limit, and you cannot proceed until you have applied a pre-filter:

This prevents Pentana from trying to draw an overly-large and complex chart.

Once you have applied some pre-filters, you will bring the volume of data down to a manageable level.  Note that Pentana may still warn you that what you are attempting to do is time-consuming:

If you answer no then Pentana does nothing, and you can apply further pre-filters.  If you click Yes then it will build the chart, and you can then click individual bars to drill down into the data per audit.

The thresholds for the Chart Point Limit and Chart Point Warning are set at Admin > System Settings > Chart:

The default settings are:

• Warning:  50.  If a chart includes more than 50 points, you will receive a warning and be able to continue
• Limit:  200.  If a chart includes more than 200 points, you will not be able to continue until pre-filters are applied.

You can change these values, but increasing to too high a value can cause performance problems.   Pentana will attempt to build the chart but will appear as Not Responding while doing so, and only be usable again once the chart is finished.

Changing the defaults

If you find that accessing the Analysis screen and selecting Actions from the dropdown is causing an unacceptable delay while the chart is drawn, then the recommended step is to adjust the Chart Point Warning to a lower value.  Then, when you access the screen and are given the option to continue you can click No, and then apply one or more pre-filters. 

For convenience, you can save the filters as a new View (and make default if required) so that you enter the screen on the filtered view.

You will then review the data set in smaller sets, by changing pre-filters as you work through.

Filtering: pre-filters vs. grid filters

It is worth bearing in mind a major difference between pre-filters, e.g.:

… and the on-grid filters, e.g.:

• Pre-filters are applied before Pentana accesses data from the server, thus limiting what is downloaded
• The data grid filters act on data that has already been brought down from the server, so you are downloading data that is then hidden

This means that it is always quicker and more efficient to use a pre-filter to limit the data that you are viewing on screen, because the server, network and client PC have to do less work.

About the error Chart Query contains errors

In order to display the Chart Point Limit message, Pentana first queries the database to calculate the number of points that would be displayed.  If the data volume is sufficiently large then:

• You will see just the spinner on the screen, no dialog box will appear
• Eventually an error will be displayed:  Chart Query contains errors
• The service error log will show:
  
  Exception of type: TngException[Executor] - Message: Msg -2, Level 11, State 0, Line 0: Execution Timeout Expired. The timeout period elapsed prior to completion of the operation or the server is not responding.

If this problem occurs:

•  Ensure that the database has been reindexed, in order to ensure that data is being retrieved as quickly as possible
• If that doesn't help, increase the command timeout so that the service waits for longer before giving up (contact Ideagen Support for more information on this)

If the data volume is this great, then going onto use saved views will be particularly important for a positive user experience.