如何在 BigQuery 中连接两个不同数据类型的数组？

Posted 2023-03-25

【中文标题】如何在 BigQuery 中连接两个不同数据类型的数组？

【英文标题】：How do I concatenate two arrays of different datatypes in BigQuery?

【发布时间】：2019-01-25 23:09:48

【问题描述】：

当我尝试连接两个相同数据类型的数组时，它工作得非常好。

一个例子如下所示

#standardSQL
WITH
  table1 AS (
  SELECT 'a' id UNION ALL
  SELECT 'b' UNION ALL
  SELECT 'c'
  ),
  table2 AS (
  SELECT 'a' id, [1,2,3,4,5] array_1 UNION ALL
  SELECT 'b', [1,2,3,4,5] UNION ALL
  SELECT 'c', [1,2,3,4,5]
  ),
  table3 AS (
  SELECT 'a' id, [10,20,30,40,50] array_2 UNION ALL
  SELECT 'b', [10,20,30,40,50] UNION ALL
  SELECT 'c', []
  ),
  joined_table as (
  select table1.id,
  table2.array_1,
  table3.array_2
  from 
    table1
  JOIN table2
  USING(id)
  JOIN table3
  using(id)
  )
SELECT
  joined_table.*,
  ARRAY_CONCAT(IFNULL(joined_table.array_1,[]), IFNULL(joined_table.array_2,[])) as concatanated_arrays
  FROM joined_table

但是当我尝试进行连接时，数组的数据类型不同，如下面的示例查询：

#standardSQL
WITH
  table1 AS (
  SELECT 'a' id UNION ALL
  SELECT 'b' UNION ALL
  SELECT 'c'
  ),
  table2 AS (
  SELECT 'a' id, [1,2,3,4,5] array_1 UNION ALL
  SELECT 'b', [1,2,3,4,5] UNION ALL
  SELECT 'c', [1,2,3,4,5]
  ),
  table3 AS (
  SELECT 'a' id, ['10','20','30','40','50'] array_2 UNION ALL
  SELECT 'b', ['10','20','30','40','50'] UNION ALL
  SELECT 'c', []
  ),
  joined_table as (
  select table1.id,
  table2.array_1,
  table3.array_2
  from 
    table1
  JOIN table2
  USING(id)
  JOIN table3
  using(id)
  )
SELECT
  joined_table.*,
  ARRAY_CONCAT(IFNULL(joined_table.array_1,[]), IFNULL(joined_table.array_2,[])) as concatanated_arrays
  FROM joined_table

我的查询没有运行并显示如下错误：

Error: No matching signature for function ARRAY_CONCAT for argument types: ARRAY<INT64>, ARRAY<STRING>. Supported signature: ARRAY_CONCAT(ARRAY, [ARRAY, ...]) at [31:3]

如何连接这两个不同类型的数组？

【参考方案1】：

无法在 BigQuery 中连接具有不同数据类型的数组。连接数组时需要相同的数据类型。

解决手头问题的一种方法是将整数数组转换为字符串，然后将它们连接起来得到组合数组。

可以在 BigQuery 标准 SQL 中使用以下查询来演示一个这样的示例：

#standardSQL
WITH
  table1 AS (
  SELECT 'a' id UNION ALL
  SELECT 'b' UNION ALL
  SELECT 'c'
  ),
  table2 AS (
  SELECT 'a' id, [1,2,3,4,5] array_1 UNION ALL
  SELECT 'b', [1,2,3,4,5] UNION ALL
  SELECT 'c', [1,2,3,4,5]
  ),
  table3 AS (
  SELECT 'a' id, ['10','20','30','40','50'] array_2 UNION ALL
  SELECT 'b', ['10','20','30','40','50'] UNION ALL
  SELECT 'c', []
  ),
  joined_table as (
  select table1.id,
  table2.array_1,
  table3.array_2
  from 
    table1
  JOIN table2
  USING(id)
  JOIN table3
  using(id)
  )
SELECT
  joined_table.*,
  ARRAY_CONCAT(IFNULL(ARRAY(SELECT CAST(value as string) FROM UNNEST(joined_table.array_1) value),[]), IFNULL(joined_table.array_2,[])) as concatanated_arrays
  FROM joined_table

但是结果最后的concatenated_arrays变成了一个重复的字符串字段。

【参考方案2】：

替代（更轻量）版本 - BigQuery 标准 SQL

#standardSQL
SELECT id, 
  ARRAY(
    SELECT CAST(val AS STRING) FROM t.array_1 val UNION ALL 
    SELECT * FROM t.array_2
  ) concatanated_arrays
FROM joined_table t   

如果应用于有问题的样本数据 - 产生预期结果