Question

我有一些数据定期加载到BigQuery数据库中。每行代表执行订单的一部分库存的移动。

相关的订单和“ order_product”信息将作为嵌套记录嵌入行中。

下面是一些行的示例：

|----------|----------|----------|------------------|------------------------|---------------------|---------------------------------|
| move_id  | quantity | order.id | order_product.id | order_product.quantity | order_product.price |      item_reference_number      |
|----------|----------|----------|------------------|------------------------|---------------------|---------------------------------|
|    1     |    1     |     1    |         1        |            1           |           5         |               ABC               |
|----------|----------|----------|------------------|------------------------|---------------------|---------------------------------|
|    2     |    1     |     1    |         2        |            1           |           7         |               DEF               |
|----------|----------|----------|------------------|------------------------|---------------------|---------------------------------|
|    3     |    1     |     1    |         2        |            1           |           7         |               XYZ               |
|----------|----------|----------|------------------|------------------------|---------------------|---------------------------------|

如您所见，该表显示了三个库存移动。全部与订单1相关。

订单1由一个订单产品1和一个订单产品2组成。订单产品1由一个库存移动组成，物料ABC的move_id 1。

Order_product 2由两个库存移动组成，项目DEF的move_id 2和项目XYZ的move_id 3。

如何编写查询以将该数据转换为具有正确的嵌套/重复字段的表？换句话说，我希望数据看起来像这样：

|-----------|------------------|------------------------|---------------------|------------------------|---------------------|----------------------------------|
| order_id  | order_product.id | order_product.quantity | order_product.price |      stock_move.id     | stock_move.quantity | stock_move.item_reference_number |
|-----------|------------------|------------------------|---------------------|------------------------|---------------------|----------------------------------|
|     1     |         1        |            1           |          5          |            1           |           1         |                ABC               |
|           |------------------|------------------------|---------------------|------------------------|---------------------|----------------------------------|
|           |         2        |            1           |          7          |            2           |           1         |                DEF               |
|           |                  |                        |                     |------------------------|---------------------|----------------------------------|
|           |                  |                        |                     |            3           |           1         |                XYZ               |
|-----------|------------------|------------------------|---------------------|------------------------|---------------------|----------------------------------|

我一直在阅读this post，这似乎表明ARRAY_AGG可能是我所需要的，但我不知道如何正确使用它来解决我的问题。

我认为我的问题是我正在努力将嵌套的order_products每行减少到一排，同时为每种order_product填充正确的嵌套/重复库存移动。

是否有可能按照我的要求去做？我非常感谢您为我指明正确的方向提供了帮助。

Answer 1

以下是用于BigQuery标准SQL

#standardSQL
SELECT order_id,
  ARRAY_AGG(product ORDER BY product.id) order_product,
  ARRAY_CONCAT_AGG(stock_move) stock_move
FROM (
  SELECT order_id, 
    STRUCT(order_product.id, order_product.quantity, order_product.price) product,
    ARRAY_AGG(STRUCT(move_id AS id, quantity AS quantity, item_reference_number AS item_reference_number)) stock_move
  FROM `project.dataset.table`
  GROUP BY order_id, order_product.id, order_product.quantity, order_product.price
  ORDER BY order_product.id -- <-- this is to make sure stock_move array is ordered as in your expected output  - but really not needed here
)
GROUP BY order_id

应用于样本数据时-上面的结果如下

我不确定这是否就是您的意思，因为您的示例仍然有点模棱两可，但是希望这能给您一个想法

还请注意：我假设您的示例中的order.id实际上是order_id，否则就没有多大意义，但是我对此可能是错误的（因为我提到您的示例仍然模棱两可”一点”）

Answer 2

以下SQL能否满足您的期望？

我以order_product方式创建了stock_move。

WITH original_table AS (
    SELECT 1 AS move_id, 1 AS quantity, STRUCT(1 AS id) AS `order`, STRUCT(1 AS id, 1 AS quantity, 5 AS price) AS order_product, "ABC" AS item_reference_number
    UNION ALL
    SELECT 2 AS move_id, 1 AS quantity, STRUCT(1 AS id) AS `order`, STRUCT(2 AS id, 1 AS quantity, 7 AS price) AS order_product, "DEF" AS item_reference_number
    UNION ALL
    SELECT 3 AS move_id, 1 AS quantity, STRUCT(1 AS id) AS `order`, STRUCT(2 AS id, 1 AS quantity, 7 AS price) AS order_product, "XYZ" AS item_reference_number
),

t1 AS (
    SELECT DISTINCT
        move_id,
        quantity,
        `order`.id AS order_id,
        order_product.id AS order_product_id,
        order_product.quantity AS order_product_quantity,
        order_product.price AS order_product_price,
        item_reference_number
    FROM original_table
),

t2 AS (
    SELECT
        order_id,
        order_product_id,
        order_product_quantity,
        order_product_price,
        ARRAY_AGG(STRUCT(move_id, quantity, item_reference_number) ORDER BY move_id) AS stock_move
    FROM t1
    GROUP BY order_id, order_product_id, order_product_quantity, order_product_price
),

t3 AS (
    SELECT
        order_id,
        ARRAY_AGG(STRUCT(order_product_id AS id, order_product_quantity AS quantity, order_product_price AS price, stock_move) ORDER BY order_product_id) AS order_product
    FROM t2
    GROUP BY order_id
)

SELECT * FROM t3

|-----------|------------------|------------------------|---------------------|----------------------------------|-----------------------------------|------------------------------------------------|
| order_id  | order_product.id | order_product.quantity | order_product.price | order_product.stock_move.move_id | order_product.stock_move.quantity | order_product.stock_move.item_reference_number |
|-----------|------------------|------------------------|---------------------|----------------------------------|-----------------------------------|------------------------------------------------|
|     1     |         1        |            1           |          5          |              1                   |           1                       |                ABC                             |
|           |------------------|------------------------|---------------------|----------------------------------|-----------------------------------|------------------------------------------------|
|           |         2        |            1           |          7          |              2                   |           1                       |                DEF                             |
|           |                  |                        |                     |----------------------------------|-----------------------------------|------------------------------------------------|
|           |                  |                        |                     |              3                   |           1                       |                XYZ                             |
|-----------|------------------|------------------------|---------------------|----------------------------------|-----------------------------------|------------------------------------------------|

如何查询非规范化BigQuery表以输出嵌套和重复字段

2 个答案: