Bigquery Standard Sql 等效于 Legacy Sql 中的 GROUP_CONCAT_UNQUOTED

Question

我目前正在使用 Bigquery 动态生成查询，使用“GROUP_CONCAT_UNQUOTED”函数将多个字符串连接为一个。

例子：

SELECT 'SELECT ID, AID, ' + 
   GROUP_CONCAT_UNQUOTED(
      'MAX(IF(KEY = "' + KEY + '", VALUE, NULL)) as [' + KEY + ']'
   ) 
   + ' FROM [project:dataset.tbl] GROUP BY 1,2 ORDER BY 1,2'
FROM (
  SELECT KEY 
  FROM [project:dataset.tbl]
  GROUP BY KEY
  ORDER BY KEY
) 

上面的查询产生另一个查询，如下所示：

SELECT 
    ID, 
    AID, 
    MAX(IF(KEY = "key1", VALUE, NULL)) as [key1],
    MAX(IF(KEY = "key2", VALUE, NULL)) as [key2],
    MAX(IF(KEY = "key3", VALUE, NULL)) as [key3] 
FROM [project:dataset.tbl]
GROUP BY 1,2 
ORDER BY 1,2

我想使用标准 SQL 而不是旧版 SQL 来做同样的事情。

Google 文档提到 STRING_AGG() 作为替代方案，但我没有得到所需的输出

Answer 1

以下是 BigQuery 标准 SQL

#standardSQL
SELECT '''
SELECT ID, AID, ''' || (
  SELECT STRING_AGG(DISTINCT "MAX(IF(key = '" || key || "', value, NULL)) AS " || key)
  FROM `project.dataset.table`
) || '''  
FROM `project.dataset.table`
GROUP BY 1,2
ORDER BY 1,2
'''

您也可以使用EXECUTE IMMEDIATE 一次性实际执行刚刚构建的查询

EXECUTE IMMEDIATE '''
SELECT ID, AID, ''' || (
  SELECT STRING_AGG(DISTINCT "MAX(IF(key = '" || key || "', value, NULL)) AS " || key)
  FROM `project.dataset.table`
) || '''  
FROM `project.dataset.table`
GROUP BY 1,2
ORDER BY 1,2
'''