Question

我在大型表（〜60GB）上运行以下形式的查询。我认为WITH OFFSET是某种形式的ORDER BY？欢迎任何想法来解决这个问题。

我收到以下错误：作业xxx失败，出现错误RESOURCE_EXHAUSTED：查询执行期间超出了资源：无法在分配的内存中执行查询。高峰使用：限制的112％。内存使用大户：ORDER BY操作：98％其他/未分配：2％；职位编号：xxx

WITH
  test_data AS (
  SELECT
    CAST(CURRENT_TIMESTAMP() as DATETIME) datetime,
    'xyz' AS grp,
    '["7f9f98fh9g4ef393d3h5", "chg3g33f26949hg6067d", "g477e5973ec04g7c3232", "0de1ec83304d761he786", "3c1h1f153530g90g35c2", "946637g145h48322686f"]' AS ids
  UNION ALL
  SELECT
    CAST(CURRENT_TIMESTAMP() as DATETIME) datetime,
    'abc' AS grp,
    '["7f9f98fh9g4ef393d3h5", "chg3g33fdsfsdfs49hg6067d", "g477e5973ec04g7c3232", "0de1ec83304d761he786", "3c1h1f153530g90g35c2", "946637g145h48322686f"]' AS ids ),

  as_list AS (
    SELECT
      datetime,
      grp,
      id,
      pos
    FROM
     test_data, UNNEST(SPLIT(REGEXP_REPLACE(JSON_EXTRACT(ids,'$'), r'[\[\]\"]', ''), ',')) AS id WITH OFFSET AS pos)

SELECT
  *
FROM
  as_list
ORDER BY
  grp, datetime

Answer 1

要解决该问题，只需删除ORDER BY或添加LIMIT

查询失败，错误为RESOURCE_EXHAUSTED（ORDER BY而不是ORDER BY）

1 个答案: