7.4 Calculation
The Calculation section (section 4 of the analysis form) defines what the RT analysis computes and where it stores the results. It contains four parts: Apply Calculation On, Rollup On Window, Output Timestamp, and Output Attributes.
7.4.1 Apply Calculation On
Apply Calculation On determines the data scope for the calculation — whether the calculation runs on the current element's own attributes or aggregates data across its child elements.
| Option | Description |
|---|---|
| Element Self | The calculation runs on the element's own attributes. Suitable for calculations on individual devices or measurement points. |
| Child Elements Aggregation | The calculation aggregates data across the element's child elements that share a common template. Suitable for cross-element summary metrics such as "compute the average power output across all turbines under this wind farm". |
When Child Elements Aggregation is selected, two additional fields appear:
| Field | Description |
|---|---|
| Child Element Template | The template that the child elements must match. Only children with this template participate in the aggregation. Automatically pre-populated if all children share the same template. |
| Subtable Filter | An optional filter expression to narrow which child elements are included in the aggregation. For example, filter to only children in a specific operating state. |
Child Elements Aggregation is only available when the element contains child elements. On leaf elements, this option is disabled and only Element Self is available.
7.4.2 Rollup On Window
The Rollup On Window checkbox (enabled by default) controls whether the calculation is aggregated over a time window.
When enabled, the Interval field specifies the length of the aggregation window (number + time unit). For example, an interval of 1 hour means each trigger firing computes the aggregate over the past 1 hour of data.
When disabled, the calculation runs over individual data points without windowed aggregation — suitable for row-level calculations or transformations.
7.4.3 Output Timestamp
The Output Timestamp dropdown specifies which timestamp is written to the output attribute for each result row:
| Option | Description |
|---|---|
| Window Start | The timestamp of the beginning of the window |
| Window End | The timestamp of the end of the window (default) |
| Timestamp | The raw timestamp (_c0), i.e. the data ingestion time. Suited to row-wise streaming output scenarios such as period triggers and anomaly detection, preserving the true ingestion time of each result row |
When Timestamp (_c0) is selected, the system automatically decides the output form based on the trigger type: data-window aggregations (sliding window with interval, count window, session window, etc.) output the aggregated moment last(_c0); row-wise streaming calculations (period triggers, anomaly detection, sliding without interval, etc.) output the raw _c0 directly.
The Offset field adds a time offset (number + unit, default 0 seconds) to the selected window boundary. This can be useful to shift the output timestamp for display alignment.
7.4.4 Output Attributes
The Output Attributes table maps calculation expressions to element attributes (and optionally event attributes when event generation is enabled).
Each row in the table has the following columns:
| Column | Description |
|---|---|
| Expression | A calculation expression. Click the cell to open the Expression Editor (see Section 3.2.10). |
| Target Element | The element where this row's result is saved. Defaults to the current element and can be switched to another element — the result is written to the target element's attribute, useful for persisting aggregation/statistics results onto summary elements. Analysis templates do not support cross-element saving (the column is hidden). |
| Element Attribute | The element attribute where the computed result is stored as a new time-series value (following the row's selected target element) |
| Event Attribute | (Visible only when event generation is enabled in section 3) An event attribute to capture the computed value at the moment the event fires |
Use the + button at the bottom of the table to add additional output rows. Each row corresponds to an independent expression, supporting the computation of multiple metrics in a single RT analysis with results written to different attributes (each row may target a different element).
Expressions throughout the analysis (output attribute expressions, trigger pre-filters, event window conditions, etc.) can reference attribute child attributes and the sub-attributes of KPI attributes. In the expression editor's attribute tree, expand the parent attribute (or the KPI attribute) and click a sub-attribute node to insert a reference of the form ${attributes['ParentName']|childAttributes['ChildName']}. A KPI attribute holds no value itself and cannot be referenced directly — you must select one of its sub-attributes.
When a KPI sub-attribute that has finished building is referenced, the computation follows the KPI data in real time; when a building-state or derived monthly/quarterly KPI sub-attribute is referenced, the stream reads the tag snapshot value (not the precise derived aggregate) — for precise derived aggregation values, use a KPI panel or query the sub-attribute directly. When an expression references a KPI sub-attribute, the UI shows a data-access semantics hint next to the expression. When a formula-type child attribute is referenced, its formula is inlined into the generated SQL (same behavior as panels) — no materialized column is required.
7.4.5 Trigger Window Pseudo Columns
The stream trigger window pseudo columns are only effective during stream computation. They are not evaluated during expression evaluation and return empty values:
| Pseudo Column | Meaning |
|---|---|
_twstart | Start time of the trigger window |
_twend | End time of the trigger window |
_twduration | Duration of the trigger window |
In the Expression Editor for Output Attributes, the left tree provides a Window Pseudo Columns category (above Substitution Parameters): expand it and click a pseudo column node to insert it. Typing in the expression input also shows matching completion suggestions.
7.4.6 Output Filter
The Output Filter is an optional condition expression that is evaluated after the RT analysis completes its calculation and before the result is written to the output attributes. The calculated result is written to the configured element attributes only when the filter evaluates to true.
Use this when you want to persist only results that satisfy a business condition, such as windows with valid samples, values outside an acceptable range, or outputs produced under a specific state. If the filter condition is not met, the trigger is still considered processed, but no value is written to the output attributes.
