标准SQL中的BigQuery Join数组

Question

我正在使用标准SQL，并且具有表Order ：

“订单”表

我试图将它与表MenuItem

“ MenuItem”表

在Order item_ids数组和MenuItem __id__整数列上，并获取MenuItem价格的数组，但出现错误：

除非引用可以取消关联，否则不支持引用其他表的关联子查询，例如将它们转换为有效的JOIN。

如何避免这个错误？

查询：

WITH menu_items AS 
( 
    SELECT
        __id__,
        price
    FROM
        `potykion.MenuItem`
)
SELECT
    *, 
    ARRAY(
        SELECT 
           price
        FROM 
           UNNEST(item_ids) AS id
        JOIN 
            menu_items 
        ON 
            id = menu_items.__id__
    ) 
FROM 
    `potykion.Order`

Answer 1

请尝试以下（BigQuery标准SQL）

WITH Orders AS (
  SELECT 1 AS id, ARRAY[1,2,3] AS item_ids UNION ALL 
  SELECT 2 AS id, ARRAY[4,5] AS item_ids UNION ALL 
  SELECT 3 AS id, ARRAY[1,4,6] AS item_ids 
),
MenuItems AS (
  SELECT 1 AS __id__, 1.1 AS price UNION ALL
  SELECT 2 AS __id__, 1.2 AS price UNION ALL
  SELECT 3 AS __id__, 1.3 AS price UNION ALL
  SELECT 4 AS __id__, 1.4 AS price UNION ALL
  SELECT 5 AS __id__, 1.5 AS price UNION ALL
  SELECT 6 AS __id__, 1.6 AS price UNION ALL
  SELECT 7 AS __id__, 1.7 AS price 
)
SELECT 
  *, 
  ARRAY(
    SELECT price 
    FROM UNNEST(item_ids) AS id 
    JOIN MenuItems 
    ON __id__ = id
  ) AS prices
FROM Orders  

表顺序：

表格菜单项：

结果：

Answer 2

在数组创建表达式中使用join的解决方案是正确的，但是不适用于单独的表。 替代解决方案是数组聚合：

WITH Orders AS (
  SELECT 1 AS id, ARRAY[1,2,3] AS item_ids UNION ALL 
  SELECT 2 AS id, ARRAY[4,5] AS item_ids UNION ALL 
  SELECT 3 AS id, ARRAY[1,4,6] AS item_ids 
),
MenuItems AS (
  SELECT 1 AS __id__, 1.1 AS price UNION ALL
  SELECT 2 AS __id__, 1.2 AS price UNION ALL
  SELECT 3 AS __id__, 1.3 AS price UNION ALL
  SELECT 4 AS __id__, 1.4 AS price UNION ALL
  SELECT 5 AS __id__, 1.5 AS price UNION ALL
  SELECT 6 AS __id__, 1.6 AS price UNION ALL
  SELECT 7 AS __id__, 1.7 AS price 
)
SELECT 
  id, ARRAY_AGG(price)
FROM Orders  
  JOIN MenuItems ON __id__ in UNNEST(item_ids)
  GROUP BY id