---

{ “title”: “HLL_RAW_AGG”, “language”: “en” }

Description

The HLL_RAW_AGG function is an aggregate function mainly used to merge multiple HyperLogLog data structures into one.

Alias

• HLL_UNION

Syntax

HLL_RAW_AGG(<hll>)
HLL_UNION(<hll>)

Parameters

Parameter	Description
<hll>	The expression to be calculated, type HLL supported.

Return Value

Returns the aggregated HLL type. If there is no valid data in the group, returns HLL_EMPTY.

Example

-- setup
create table test_uv(
    id int,
    uv_set string
) distributed by hash(id) buckets 1
properties ("replication_num"="1");
insert into test_uv values
    (1, ('a')),
    (1, ('b')),
    (2, ('c')),
    (2, ('d')),
    (3, null);

select HLL_CARDINALITY(HLL_RAW_AGG(hll_hash(uv_set))) from test_uv;

+------------------------------------------------+
| HLL_CARDINALITY(HLL_RAW_AGG(hll_hash(uv_set))) |
+------------------------------------------------+
|                                              4 |
+------------------------------------------------+

select HLL_CARDINALITY(HLL_RAW_AGG(hll_hash(uv_set))) from test_uv where uv_set is null;

+------------------------------------------------+
| HLL_CARDINALITY(HLL_RAW_AGG(hll_hash(uv_set))) |
+------------------------------------------------+
|                                              0 |
+------------------------------------------------+