RollingMin / Computation Layer

Calculate the rolling (moving) minimum over a sliding window. Similar to pandas.rolling().min() or R's roll_min().

Mathematical form: $y_{i} = j = i - w + 1 min i (x_{j} \cdot w_{j})$ where $w$ is the window size and $w_{j}$ are optional weights.

Common applications:

Stock market analysis (support levels)
Resource optimization (minimum requirements)
Quality control (minimum thresholds)
Risk analysis (worst performance)
System monitoring (resource bottlenecks)
Environmental monitoring (minimum levels)
Performance benchmarking

Note: Useful for identifying lower bounds and support levels in data.

Table

Select

column

The numeric column to compute rolling minimums for. Typical applications:

Asset prices (support levels)
Resource utilization (minimum capacity)
Temperature readings (frost warnings)
Performance metrics (worst case analysis)
Water levels (drought monitoring)

WindowSize

u32

Number of observations in each sliding window. Common periods:

24: Daily minimum (hourly data)
7: Weekly minimum
30: Monthly minimum
252: Annual minimum (trading days)

Larger windows capture longer-term floor levels.

Weights

[f64, ...]

Optional weights for window values. Must match window size if provided. Note:

Affects minimum selection when comparing values
Can prioritize recent or specific positions
Less common in minimum calculations than means

Consider whether weighted minimums align with analysis goals.

MinPeriods

oneof

AsWindow

Require complete windows for minimum calculation. Ensures all values in window are considered when identifying minimum. Critical for accurate lower bound detection and threshold monitoring.

Period

Value

u32

Minimum number of valid observations (Value) required for minimum calculation. Examples:

1: Find minimum of any available values
window_size/2: Require majority of window
window_size: Require all values in window Controls trade-off between data availability and reliable minimum detection.

Center

bool

false

Window label position. When true, aligns to window center; when false, to end. Effects:

true: Better for pattern analysis and historical minimums
false: Preferred for real-time minimum monitoring

Name for the new column. If not provided, the system generates a unique name. If AsColumn matches an existing column, the existing column is replaced. The name should follow valid column naming conventions.