Quantile digest functions

Data structures

A quantile digest is a data sketch which stores approximate percentile information.

Trino Type	Parameter Options	Description
qdigest	bigint, double, real	Represents the set of numbers that may be ingested by the qdigest. They may be merged without losing precision. For storage and retrieval, they may be cast to/from VARBINARY seamlessly.

Functions

merge()

Function	Description	Return Value
merge(qdigest)	Merges all input qdigests into a single qdigest.	qdigest

value_at_quantile()

Function	Description	Return Value
value_at_quantile(qdigest(T), quantile)	Returns the approximate percentile value from the quantile digest given the number quantile.	Numeric

quantile_at_value()

Function	Description	Return Value
quantile_at_value(qdigest(T), T)	Returns the approximate quantile number between 0 and 1 from the quantile digest given an input value.	Numeric

values_at_quantiles(qdigest(T), quantiles)

Function	Description	Return Value
values_at_quantiles(qdigest(T), quantiles)	Returns the approximate percentile values as an array given the input quantile digest and array of values between 0 and 1 representing quantiles.	Array(T)

qdigest_agg()

Function	Description	Return Value
qdigest_agg(x) -> qdigest([same as x])	Returns the qdigest composed of all input values of x.	qdigest([same as x])
qdigest_agg(x, w) -> qdigest([same as x])	Returns the qdigest composed of all input values of x using the per-item weight w.	qdigest([same as x])
qdigest_agg(x, w, accuracy) -> qdigest([same as x])	Returns the qdigest composed of all input values of x using the per-item weight w and maximum error of accuracy.	qdigest([same as x])