APPROX_COUNT_DISTINCT
description
Syntax
APPROX_COUNT_DISTINCT(expr)
返回类似于 COUNT(DISTINCT col) 结果的近似值聚合函数。
它比 COUNT 和 DISTINCT 组合的速度更快,并使用固定大小的内存,因此对于高基数的列可以使用更少的内存。
example
MySQL > select approx_count_distinct(query_id) from log_statis group by datetime;
+-----------------+
| approx_count_distinct(`query_id`) |
+-----------------+
| 17721 |
+-----------------+
keyword
APPROX_COUNT_DISTINCT