在 BigQuery 中对串联字符串（行）进行分组

Question

我正在使用 Google BigQuery & 我有一个如下所示的查询：

 SELECT
        prod.abc
        uniqueid,
        variable2,
        cust.variable1,
        purch.variable2,
        from mydata.order
        left join
        UNNEST(purchases) as purch,
        UNNEST(codes_abs) as cod, UNNEST(cod.try_products) as prod

当我这样做时，会产生一个如下所示的表：

    |prod.abc| uniqueid | variable2 | ...|
    |APP123  | customer1| value     | ...|
    |BLU155  | customer1| value     | ...|
    |TRI134  | customer1| value     | ...|
    |LO123   | customer2| value     | ...|
    |ZU9274  | customer2| value     | ...|
    |TO134   | customer3| value     | ...|

我想做的是将“prod.abc”列中的值连接起来，按“uniqueid”对它们进行分组，并用“,”分隔它们。然而，我在网上找到了许多解决方案，因为我在查询中取消了其他变量的嵌套，所以我找到的解决方案似乎都不适合我的情况。这些值不需要以任何方式排序。基本上，我想要的结果是：

    |prod.abc                  | uniqueid | variable2 | ...|
    |APP123, BLU155, TRI134    | customer1| value     | ...|
    |LO123, ZU9274             | customer2| value     | ...|
    |TO134                     | customer3| value     | ...|

获得一个像这样的保留重复项的表也可以，因为我可以稍后删除它们：

|prod.abc                  | uniqueid | variable2 | ...|
|APP123, BLU155, TRI134    | customer1| value     | ...|
|APP123, BLU155, TRI134    | customer1| value     | ...|
|APP123, BLU155, TRI134    | customer1| value     | ...|
|LO123, ZU9274             | customer2| value     | ...|
|LO123, ZU9274             | customer2| value     | ...|
|TO134                     | customer3| value     | ...|

任何帮助深表感谢。谢谢你！

Answer 1

分别取消嵌套：聚合有效吗？

SELECT STRING_AGG(item.abc, ',')
       uniqueid, variable2, cust.variable1, purch.variable2
FROM mydata.order LEFT JOIN
     UNNEST(purchases) as purch
     ON true LEFT JOIN
     UNNEST(codes_abs) as cod
     ON true LEFT JOIN
     UNNEST(cod.try_items) as item
     ON true
GROUP BY uniqueid, variable2, cust.variable1, purch.variable2;