The example screen capture above shows a Boolean point with a COV tag (highlighted).

Consider the following analytic request and result for history data that does not include a COV tag.

Notice that this example uses an analytic trend ORD scheme (1) to identify the trend data (2). Analytic trend, analytic value, and analytic rollup are three different ORD schemes particular to the framework. It is important to understand how to use these ORD schemes, as they are the foundation of all framework functionality.

The ORD itself is:

analytictrend:data=n:history&interval=fifteenMinutes&rollup=first&timeRange=lastMonth

For clarity, it is asking for data identified by the n:history tag, using:

Based on the n:history tag, the output (2) shows that for all of last month, the point went true, first thing Tuesday morning, at midnight, and it went false at the very end of the same day. No data were collected on any other day, since the value did not change. The ORD requests an interval of 15 minutes but instead of one row every 15 minutes, the request returns the same table that is visible on the COV history. This is because the framework is not aware that the history type for this point is COV.

To resolve this issue, add a tag that identifies the history type as COV. The framework applies the NEQL query defined by the Test Cov Neql property and returns true for the point—the point’s history type is COV.

Now that the history is identified as COV, we get a 15 minute rollup (highlighted), showing the correct value for the timestamp.

All Analytic charts, tables, points, and alerts use ORD schemes. So, when something, for example, an alert that uses an algorithm, is not working the way you expect it to work, try the same algorithm using the same ORD scheme that the alert uses. This lets you see the data the alert is actually receiving as input.

Alternatively, you can use some other tag, and modify the NEQL query. You do not have to use Haystack tags for this.