Recap: aggregates

An aggregate summarises a group of values: count, sum, avg, min, max.
Other head arguments are grouping keys: one aggregate per key tuple; with no keys, aggregation is global.
An aggregate rule still derives facts: one summary fact per group.
An aggregate sees one value per row (a bag, not a set): count, sum and avg keep duplicate values, so count(*) counts rows, not distinct values.
Aggregates cannot be recursive, but often we can enumerate, then aggregate.
On the Titanic data: survival rates, average fares and ages, the youngest and oldest survivors.

Next: Part 4 steps back for the theory, the tie to SQL, and working with JSON.