Data aggregation
The connector supports aggregating data from the data source before sending it to the TrendMiner appliance. For most common historians this aggregation is done on the historian but in case a data source does not support data aggregation, the aggregation can be enabled on the connector.
Data aggregation can be configured for all data source providers at once, or for one specific data source by entering the data source name in the configuration.
Data aggregation can be configured by adding the key to the connector configuration file.
Multiple aggregation options are available.
Parameter | Description |
|---|---|
aggregation.<provider/name>.type | Aggregation algorithm: “None”, “Unique”, “Average”, “Plot” (default) |
aggregation.<provider/name>.interval.ms | Aggregation step interval in milliseconds, 1000 milliseconds by default |
aggregation.<provider/name>.stepped.enabled | Set to ‘True’ to use 'Unique' aggregation for String, Digital and Discrete tags . |
Aggregation type | Description |
|---|---|
None | Aggregation is switched off |
Unique | Non-unique values points are removed except for the first and last points of an interval |
Average | For each interval period, the average value point is provided |
Plot | For each interval period start point, end point, max value point and min value point are returned. Non-unique start and end points are removed. |
Important
Provider names used in the configuration should be the technical names and not the names visualized in the TrendMiner UI.
Find out here on how to get the technical name of a provider.
Note
ODBC providers will by default apply plot optimization based on the index resolution up to 5s index resolution (12 intervals per minute).
Examples:
<add key="aggregation.odbc.type" value="Plot"/>
All data coming from ODBC data sources via this connector will return plot optimized data to the appliance.
<add key="aggregation.odbc.interval.ms" value="60000"/>
All ODBC data source data will be aggregated over 1 minute intervals.
<add key="aggregation.plantX.stepped.enabled" value="true"/>
All String, Digital and Discrete tags from data source with name “plantX” will only return unique values per interval.
Supported as of connector version: <3.1.0>
Supported for all available data source providers.