Skip to main content
What this builds. Four rows of vendor spend, rolled up with aggregate(...) into totals, averages, and unique-vendor counts — whole-table and filtered to a single region. Then a quick look at the run / run_status shape, which is how you fill Pipeline- and Agent-generated columns in bulk. You’ll end up with. A working call to aggregate for the common reductions, plus a feel for the task handle run returns. The example uses a plain table (no generator columns) so run is a no-op here — see columns-full for the wiring that makes run do real work.

Expected output

The aggregation result map is keyed by column name, so passing two aggregations on the same column in a single aggregate(...) call returns only one of them. When you need several reductions on the same column, issue one call per (column, type) pair.

See also

Generators in the overview

Bind a column to a Pipeline or Agent so run does real work.

AggregationType reference

All 21 supported reductions and their column-type requirements.

End-to-end walkthrough

Aggregations inside a vendor-scorecard workflow.