Quantiles (DoublesSketch) Functions

Classic quantiles sketch (DoublesSketch) — the original DataSketches quantile algorithm. Provides the same functionality as KLL but with different trade-offs. For new projects, prefer KLL which is more space-efficient.

Table of contents

1. Aggregation Functions
   1. ds_quantiles_sketch
   2. ds_quantiles_union
2. Scalar Functions
   1. ds_quantiles_quantile
   2. ds_quantiles_quantiles
   3. ds_quantiles_rank
   4. ds_quantiles_cdf
   5. ds_quantiles_pmf
   6. ds_quantiles_n
   7. ds_quantiles_k
   8. ds_quantiles_stringify
3. KLL vs Quantiles

---

Aggregation Functions

ds_quantiles_sketch

ds_quantiles_sketch(value) -> varbinary
ds_quantiles_sketch(value, k) -> varbinary

Parameter	Type	Description
value	DOUBLE, REAL, or BIGINT	Numeric values
k	INTEGER	Accuracy parameter (default 128)

ds_quantiles_union

ds_quantiles_union(sketch) -> varbinary

---

Scalar Functions

ds_quantiles_quantile

ds_quantiles_quantile(sketch, rank) -> double

ds_quantiles_quantiles

ds_quantiles_quantiles(sketch, ranks) -> array(double)

ds_quantiles_rank

ds_quantiles_rank(sketch, value) -> double

ds_quantiles_cdf

ds_quantiles_cdf(sketch, splitPoints) -> array(double)

ds_quantiles_pmf

ds_quantiles_pmf(sketch, splitPoints) -> array(double)

ds_quantiles_n

ds_quantiles_n(sketch) -> bigint

ds_quantiles_k

ds_quantiles_k(sketch) -> bigint

ds_quantiles_stringify

ds_quantiles_stringify(sketch) -> varchar

---

KLL vs Quantiles

Property	KLL	Quantiles (DoublesSketch)
Space efficiency	Better	Larger for same accuracy
Merge speed	Similar	Similar
Algorithm	Near-optimal	Classic
Recommendation	Preferred	Legacy compatibility

All functions behave identically to their ds_kll_* counterparts. See KLL Sketch for detailed parameter descriptions and examples.