Approximate aggregate functions are functions that produce approximate results,
and use fewer resources than the equivalent exact aggregation. For example,
approx_count_distinct
can be used in place of COUNT (DISTINCT ...)
to
significantly reduce memory usage when the number of distinct inputs is large.
approx_count_distinct
Return an estimated number of distinct, non-NULL values in the input. This is an
approximate version of COUNT(DISTINCT ...)
.
Internally, this uses a HyperLogLog sketch and yields roughly a 1.6% relative error.
approx_quantile
Computes a value for the given quantile such that approximately count(input) * quantile
numbers are smaller than the returned value.
Internally uses a T-Digest data sketch.
approx_unique
Alias of approx_count_distinct
Return an estimated number of distinct, non-NULL values in the input. This is an
approximate version of COUNT(DISTINCT ...)
.
Internally, this uses a HyperLogLog sketch and yields roughly a 1.6% relative error.