Question

我是大问题的新手，需要创建烫发。日志文件和引用表中的表。我是新手，并做错了什么，但无法弄清楚原因。请帮忙。

日志文件（示例）

-Event_Time，USER_ID，类型

-1 / 1 / 2017,123_abc，一个

-1 / 2 / 2017,123_bcd，B

-1 / 2 / 2017,123_abc，C

参考表（示例）

-Type Partner

-a 1

-b 2

-c 3

-d 3

create table workarea.SummaryTable AS (
User_ID string,
TotalCount integer,
imps_time timestamp,
Partner integer)

insert into workarea.SummaryTable
select distinct User_ID,
COUNT(*) as TotalCount,
MIN(TIME) as imps_time,
SUM(case when Partner = '1' then 1 else 0 end) as 1,
SUM(case when Partner = '2' then 1 else 0 end) as 2,
SUM(case when Partner = '3' then 1 else 0 end) as 3

from workarea.logfile i
join workarea.referencetable r on i.Type=r.Type
where CID=10848805 
group by USER_ID

Answer 1

..并做错了什么但无法弄清楚为什么

到目前为止我已经确定的失败点

CREATE TABLE语句在BigQuery中不可用 Data Manipulation Language仅允许您INSERT，DELETE和UPDATE

您需要拥有表pre-created才能使用/插入数据

别名不能以数字开头 -

所以下面的片段不正确

SUM(CASE WHEN Partner = '1' THEN 1 ELSE 0 END) AS 1,    
SUM(CASE WHEN Partner = '2' THEN 1 ELSE 0 END) AS 2,    
SUM(CASE WHEN Partner = '3' THEN 1 ELSE 0 END) AS 3

你应该使用像

SUM(CASE WHEN Partner = '1' THEN 1 ELSE 0 END) AS Partner_1,    
SUM(CASE WHEN Partner = '2' THEN 1 ELSE 0 END) AS Partner_2,    
SUM(CASE WHEN Partner = '3' THEN 1 ELSE 0 END) AS Partner_3

某些字段在引用表中看起来不存在，但您在最终查询中使用它们：time中的MIN(time) as imps_time和CID中的WHERE CID=10848805
目标表的模式有4个字段 - 而select语句的模式有6个字段。你应该清楚这一点!!他们必须匹配！

可能的“解决方案”（BigQuery Standard SQL）

我假设（仅为了在这里取得一些进展）目的地表的架构在现实中如下所示

User_ID STRING,
TotalCount INT64,
imps_time TIMESTAMP,
Partner_1 INT64,
Partner_2 INT64,
Partner_3 INT64

在这种情况下 - 下面的查询应该产生正确的插入结果

#standardSQL
SELECT 
  User_ID,
  COUNT(*) AS TotalCount,
  MIN(Event_Time) AS imps_time,
  SUM(CASE WHEN Partner = '1' THEN 1 ELSE 0 END) AS Partner_1,
  SUM(CASE WHEN Partner = '2' THEN 1 ELSE 0 END) AS Partner_2,
  SUM(CASE WHEN Partner = '3' THEN 1 ELSE 0 END) AS Partner_3
FROM `workarea.logfile` i
JOIN `workarea.referencetable` r ON i.Type=r.Type
-- WHERE CID=10848805 
GROUP BY USER_ID

从日志表和引用表

1 个答案: