...
A StatisticsSummary transformation will compute the configured statistic function for the input values to get one result value (the summary). This uses only the input values within the relativeViewPeriod that is defined in the timeSeriesSet of the output variable - unless the readWriteMode of the input timeseries is set to 'read complete forecast'. The result output value is stored at T0 (time zero) in the output timeSeries.
...
- count
- countFlags
- kurtosis
- max
- mean (only nonequidistant as input)
- median
- min
- percentileExceedence
- percentileNonExceedence
- quartile
- rootMeanSquareError
- rsquared
- skewness
- standardDeviation
- sum
- variance
The below Below config example counts the unreliable flags per month from a 15 minute timestep seriesidentifies the maximum of a timeseries. Note that the output timeseries is a single value at T0 only. The time step of the output timeseries must be nonequidistant.
| Code Block | ||
|---|---|---|
| ||
<transformation id="count flagsIdentify_max"> <statisticsSerial> <countFlags><statisticsSummary> <period><max> <relativePeriod unit="day" start="-120" end="0"/> <cardinalTimeStep unit="month" timeZone="GMT+1"/> <inputVariable> <extendStart>true</extendStart> <extendEnd>true</extendEnd> <variableId>P_stations_1h</variableId> </period> <inputVariable> <variableId>INPUT</variableId> </inputVariable> <reliability>non_missing_unreliable</reliability> <validationRule>any</validationRule> <outputVariable> <timeSeriesSet> <moduleInstanceId>CountFlags</moduleInstanceId> <valueType>scalar</valueType> <parameterId>Number.Unreliables</parameterId> <locationSetId>LOCSET</locationSetId> <timeSeriesType>external historical</timeSeriesType> <timeStep unit="month" timeZone="GMT+1"/> <readWriteMode>add originals</readWriteMode> </timeSeriesSet> <variableId>P_max</variableId> </outputVariable> </outputVariable>max> </countFlags>statisticsSummary> </statisticsSerial> </transformation> |