Question

我正在尝试编写一个查询，该查询使用子查询来生成具有记录数的列。我已经在代码下方和下面的表中编写了代码，当我删除子查询时，我确实生成了列消息和count_date，但无法生成terr和sat列（这些列将在子查询内）。我已经能够完成一项历史记录记录，但是当我使用的间隔大于1时就无法做到。

我需要在子查询中进行些什么更改才能得到

SELECT count(*) as TotalMessages, Date(ingestion_time) as Date_ingestion, 
  (SELECT count(*) 
    FROM myTable
    WHERE collection_type = '1' AND date(ingestion_time) >=                       
 (DATE_SUB(CURRENT_DATE(), INTERVAL 350 DAY)) AND
      date(ingestion_time) < (CURRENT_DATE())
      GROUP BY date(ingestion_time)
      ) AS terr                      
FROM 
  myTable
WHERE 
  date(ingestion_time) >= (DATE_SUB(CURRENT_DATE(), INTERVAL 350 DAY)) AND
      date(ingestion_time) < (CURRENT_DATE())
GROUP BY 2
ORDER BY 2;

下面是我的想法，对我来说，我将回去超过3个月。

Messages      | count_date | terr | sat 
500           | 2019-11-13 | 486  | 14
710           | 2009-11-12 | 700  | 10
450           | 2009-11-11 | 440  | 10
767           | 2009-11-10 | 760  | 7

and so forth until the last possible count from a day.

Answer 1

通过查看查询和预期结果，我怀疑您实际上是在寻找条件聚合。如果是这样，则不需要子查询，您只需执行条件求和，如下所示：

SELECT 
    count(*) as TotalMessages, 
    Date(ingestion_time) as Date_ingestion,
    sum(case when collection_type = '1' then 1 end) as terr,
    sum(case when not collection_type = '1' then 1 end) as sat
FROM 
  myTable
WHERE 
  date(ingestion_time) >= (DATE_SUB(CURRENT_DATE(), INTERVAL 350 DAY)) AND
      date(ingestion_time) < (CURRENT_DATE())
GROUP BY 2
ORDER BY 2;

每天获取SQL Bigquery中所有记录，select语句中的子查询的每日记录数

1 个答案: