https://pizza.rs/tags/cardinality/Recent content in Cardinality on INFINI PizzaHugoenhttps://pizza.rs/docs/references/aggregation/metric/cardinality/Mon, 01 Jan 0001 00:00:00 +0000https://pizza.rs/docs/references/aggregation/metric/cardinality/<h1 id="cardinality-aggregation"> Cardinality aggregation <a class="anchor" href="#cardinality-aggregation">#</a> </h1> <p>A single-value metrics aggregation that counts the number of distinct values of a field — how many different <code>type</code>s were sold, how many different <code>host</code>s logged, how many different authors wrote.</p> <h2 id="how-it-works"> How it works <a class="anchor" href="#how-it-works">#</a> </h2> <p>The engine counts with a <strong>HyperLogLog++</strong> sketch — tuned so that small counts stay exact:</p> <ul> <li>Values are hashed (MurmurHash3, 64-bit) into a sketch that keeps an exact set while the number of distinct values stays at or below <code>precision_threshold</code>; those counts are returned verbatim.</li> <li>Above the threshold the sketch promotes to dense HyperLogLog registers, so memory stops growing with the data: roughly 1.6% relative error at the default threshold of 3000, tightening to about 0.4% at the maximum of 40000. The returned value is always a rounded whole number.</li> <li>When documents lack the field and <code>missing</code> is set, they all count as one additional distinct value; without <code>missing</code> they are simply not counted.</li> </ul> <p>Distributed queries ship the sketch itself with each shard&rsquo;s partial result and the coordinating node <strong>unions the sketches</strong> before estimating — the same number a single-node execution over all shards' documents would produce. (Cardinality nested inside bucket sub-aggregations still merges its per-shard finalized counts by maximum, a lower bound, when no sketch travels with them.)</p>