Histories collect data either on change of value or at some fixed interval. Unless the framework filters values prior to recording history records, it is possible the history records might contain values that are anomalies, such as infinite values, values that are not numbers (NaN, Not a Number), or erroneously high or low values. Analysis of historical data might be affected by those anomalies or there may be a need to simply exclude certain records based on the record’s value or timestamp. Filter blocks in algorithms ensure that the algorithm processes only valid data.

Analytics supports five filter blocks. The following table summarizes when each block applies.