Timeseries pivot
Overview
timeseries-pivot selects named columns across matching series inputs and emits one scoped output column per source.
Example configuration
pattern: series:///sensors/* # Required: example source pattern.
columns: [temperature] # Required: example measurement columns to pivot.
time_column: timestamp # Default: source event-time column.
transforms: null # Default: no input transforms.
Use and behavior
Install with pond mknod timeseries-pivot /derived/temperatures --config-path pivot.yaml, then inspect and query the resulting view.
The first wildcard capture is the source alias; without a capture, the basename is used. No matches and duplicate aliases fail. Missing selected columns are null-padded.
Full configuration
Settings
| Field | Accepted type | Required | Default | Description |
|---|---|---|---|---|
columns | array of string | yes | not specified | List of column names to pivot across all matched inputs |
pattern | Watertown Url | yes | not specified | Pattern to match input files (e.g., "series:///combined/*") |
time_column | string | no | "timestamp" | Time column name (default: "timestamp") |
transforms | array | null | no | not specified | Optional list of table transform factory paths to apply to input TableProvider. Transforms are applied in order before SQL execution. Each transform is a path like "/etc/hydro_rename" |
Watertown Url
Accepted value: string.