计算每个相关值的每个季度的平均值

Question

我在BigQuery中有一个查询。我想知道超过四分之一的平均值。使用我当前的SQL，id1的Q1周期值对于id2是相同的。

这就是我所拥有的，价值观很好：

row|averages|quarter|identifier
-----------------------------
1  | 10     | 1     | id1
2  | 20     | 2     | id1
3  | 30     | 1     | id2
4  | 40     | 2     | id2

这是我为上面的结构编写的SQL，它给出了很好的值：

WITH
  index_cal AS (
  SELECT
    values-01,
    kind,
    EXTRACT (QUARTER  FROM  date) as QUARTER,
    date,
  FROM
    `project.dataset.table`,
  geom AS (
  SELECT
    identifier
  FROM
    `project.dataset.table2` )
SELECT
   AVG(values-01) AS averages,
    QUARTER AS quarter,
    geom. identifier as identifier
  FROM
    index_cal as g
INNER JOIN
  geom
ON
  INTERSECTS(g.kind,
    geom. identifier)
GROUP BY
  identifier
  quarter

我想要的是为每个标识符的每个季度分组值，使每个标识符只有1个关联的行：

row | averages | quarter | identifier
----------------------------------
1   | 10       | 1       | id1
    | 20       | 2       |
----------------------------------
2   | 30       | 1       | id2
    | 40       | 2       |
----------------------------------

为了获得所需的结构，使得id1只有1个关联的行，同样对于所有标识符，我编写了这个SQL查询：

WITH
  index_cal AS (
  SELECT
    values-01,
    kind,
    EXTRACT (QUARTER  FROM  date) as QUARTER,
    date,
  FROM
    `project.dataset.table`,
  geom AS (
  SELECT
    identifier
  FROM
    `project.dataset.table2` )
SELECT
  ARRAY(
  SELECT
    AS STRUCT AVG(values-01) AS averages,
    QUARTER AS quarter
  FROM
    index_cal
  GROUP BY
    QUARTER ) as INDEX,
  geom. identifier as identifier
FROM
  index_cal AS g
INNER JOIN
  geom
ON
  INTERSECTS(g.kind,
    geom. identifier)
GROUP BY
  identifier

在运行此查询时，我获得按季度分组的所有标识符的平均值，以便对所有标识符重复值（例如，在这种情况下为15和25）：

row | averages | quarter | identifier
----------------------------------
1   | 15       | 1       | id1
    | 25       | 2       |
----------------------------------
2   | 15       | 1       | id2
    | 25       | 2       |
----------------------------------
2   | 15       | 1       | id3
    | 25       | 2       |
----------------------------------

我最后想要回答的是基于values-01的季度区间identifier的平均值。目前它们对于identifier的任何值都是相同的。

Answer 1

在给出原始值的原始查询上使用ARRAY_AGG解决了它

with final_cal as (WITH
  index_cal AS (
  SELECT
    values-01,
    kind,
    EXTRACT (QUARTER  FROM  date) as QUARTER,
    date,
  FROM
    `project.dataset.table`,
  geom AS (
  SELECT
    identifier
  FROM
    `project.dataset.table2` )
SELECT
   AVG(values-01) AS averages,
    QUARTER AS quarter,
    geom. identifier as identifier
  FROM
    index_cal as g
INNER JOIN
  geom
ON
  INTERSECTS(g.kind,
    geom. identifier)
GROUP BY
  identifier
  quarter)
SELECT identifier, ARRAY_AGG(STRUCT(averages, quarter)) from final_cal GROUP BY identifier