Time Series Data Treatments

Data normalization is a preprocessing step that removes duplicate, overlapping, missing, or out-of-order entries. This process creates a consistent time series.

The normalization process aggregates and disaggregates incoming time series data to produce a uniform time series. For each field in time series data, you must specify a treatment for the aggregation.

The following list displays the options available for time series data treatments:

• PREVIOUS: See AggOp#PREVIOUS
• EARLIEST: See AggOp#EARLIEST
• LATEST: See AggOp#LATEST
• AND: See AggOp#AND
• OR: See AggOp#OR
• COUNT: See AggOp#COUNT
• ROLLINGCOUNT: See AggOp#ROLLINGCOUNT
• INTEGRAL: See AggOp#INTEGRAL
• SUM: See AggOp#SUM
• PROD: See AggOp#PROD
• RATE: See AggOp#RATE
• AVG: See AggOp#AVG
• MIN: See AggOp#MIN
• MAX: See AggOp#MAX
• STDDEV: See AggOp#STDDEV
• VARIANCE: See AggOp#VARIANCE
• PERCENTILE: See AggOp#PERCENTILE
• MEDIAN: See AggOp#MEDIAN
• ALL: See AggOp#ALL

Use the @ts notation to specify a treatment when declaring a field.

The following example defines a time series Type that uses the avg data treatment for the temperature field:

entity type CityMeasurementDataPoint mixes TimedDataPoint<CityMeasurementHeader> schema name "CITYMEASUREMENTDATAPOINT" {
   // The measured temperature in degrees Celsius
   @ts(treatment='avg')
   temperature: double
}

See AggOp Type (aggregation operation Type) for more detailed information on each treatment.