BigQuery联接基于（Array CONTAINED IN Array）条件的2个表

Posted 2021-04-29

我正在努力实现以下目标。假设我有两个表：

WITH table_1 as (
SELECT
* FROM UNNEST([
  STRUCT([1] as A, [2,3,4] as B),
  STRUCT([2],[6,7])
  ])
)

表2：

WITH example as (
SELECT
* FROM UNNEST([
  STRUCT([1,2] as C, [77] as D),
  STRUCT([3,4],[88]),
  STRUCT([4],[99])
  ])
)

我想基于以下条件合并table_1和table_2，即C的所有值都必须在B中：

SELECT A, C, D FROM table_1 LEFT JOIN table_2 ON C CONTAINED IN B

这将产生下表：

我的问题是，是否/如何能够获得预期的结果。我无法为两个数组编写CONTAINED IN语句作为LEFT JOIN语句的条件。另一个要求是表1包含1亿行，表2包含25,000行。因此，解决方案必须高效。我知道这增加了问题的难度...：P

非常感谢您的帮助！

答案

WITH table_1 as (
SELECT
* FROM UNNEST([
  STRUCT([1] as A, [2,3,4] as B),
  STRUCT([2],[6,7])
  ])
),
table_2 as (
SELECT
* FROM UNNEST([
  STRUCT([1,2] as C, [77] as D),
  STRUCT([3,4],[88]),
  STRUCT([4],[99])
  ])
)

SELECT table_1.A, table_2.C, table_2.D
FROM table_1 , table_2 , UNNEST([
                      (SELECT ARRAY_LENGTH(table_2.C) - COUNT(1) 
                      FROM UNNEST(table_2.C) AS col_c 
                      JOIN UNNEST(table_1.B)  AS col_b 
                      ON col_c = col_b)]) AS x
WHERE x = 0