hadoop - 如何在配置单元中获取复杂数据类型列的长度

Question

我有一个配置单元表包含 array 和 map 类型的列，我想过滤记录，其中 array/map 列包含超过 N 元素，该怎么做？

DDL:

create table test (id string, v1 array<int>, v2 map<string,string>)

查询:

select * from test where length(v1)>10 or length(v2)>10

Answer 1

select * from test where size(v1)>10 or size(v2)>10

---

演示

create table test (id string, v1 array<int>, v2 map<string,string>);
insert into test select 1,array(1,2,3,4,5),map('K1','V1','K2','V2','K3','V3');

---

select  size(v1),size(v2)
from    test
;

---

+----+----+
| c0 | c1 |
+----+----+
|  5 |  3 |
+----+----+