Database Histogram
An optimizer statistic that summarizes a non-uniform column-value distribution so cardinality and selectivity estimates can differ by predicate value.
When a few values dominate a column and the rest are rare, a distinct-count-only model can effectively assume the same frequency for every value. A histogram gives the optimizer more information about that skew and can improve row-count estimates for specific literals or binds.
A histogram is not an index and does not automatically make a query faster. Unnecessary or stale histograms can affect plan stability, so collection policy and workload value distribution matter. In Oracle Database and PL/SQL, this belongs to the estimation layer that precedes access-path selection.