Skip to content

rolling-min

Input Stack:
n: Int
expr: TimeSeriesExpr
Output Stack:
TimeSeriesExpr
 

Find the minimum value within a rolling window of datapoints. This operation is useful for tracking low-water marks, establishing lower bounds for noisy data, and creating envelope functions that follow the lowest points in fluctuating signals.

Parameters

  • expr: The time series expression to find rolling minimum for
  • n: Window size (number of datapoints including current value)

Behavior

  • Minimum selection: Returns the lowest value in the current window
  • Missing data: NaN values are ignored in the calculation
  • All NaN window: If all values in window are NaN, emits NaN
  • Window-based: Uses fixed number of datapoints, not time duration

Alerting Example

Using rolling min to establish recent low baseline:

name,sps,:eq,:sum,
:dup,
5,:rolling-min

This creates both the current value and its recent minimum for comparison.

Data Processing Example

Input 3,:rolling-min
0 0
1 0
-1 -1
NaN -1
0 -1
1 0
1 0
1 1
1 1
0 0

Examples

5-point rolling minimum:

BeforeAfter
name,sps,:eq,
:sum
name,sps,:eq,
:sum,
5,:rolling-min

Since: 1.6