Google bigquery BigQuery中字符串数组的差异

我试图使用标准SQL确定BigQuery中两个字符串数组之间的差异

比如说：

id    cust_id    orders     high_value_orders
1     1          [a,b,c,d]    [a,b]

我想做的是：

id    cust_id    diff
1     1          [c,d]

基本上执行设置操作

订单-高值订单

---

我已经尝试了一些东西，我知道它可能需要一些不必要的东西，但我不知道如何准确地编写它。任何帮助都将不胜感激。

以下是BigQuery标准SQL

#standardSQL
SELECT *,
  ARRAY(
    SELECT element FROM (
      SELECT DISTINCT element FROM t.orders element UNION ALL
      SELECT DISTINCT element FROM t.high_value_orders element 
    )
    GROUP BY element
    HAVING COUNT(1) = 1
  ) AS diff
FROM `project.dataset.table` t   

如果要应用于您的示例数据，请参见下面的示例

#standardSQL
WITH `project.dataset.table` AS (
  SELECT 1 id, 1 cust_id, ['a','b','c','d'] orders, ['a','b'] high_value_orders
)
SELECT *,
  ARRAY(
    SELECT element FROM (
      SELECT DISTINCT element FROM t.orders element UNION ALL
      SELECT DISTINCT element FROM t.high_value_orders element 
    )
    GROUP BY element
    HAVING COUNT(1) = 1
  ) AS diff
FROM `project.dataset.table` t   

结果是

---

谢谢@Mikhail Berlyant！