sql - Big Query 或 SQL reshape 数据

Question

我使用 bigquery 存储数据
例如我有 table

userId|event  |count
------------- |
1     |event1 |1
1     |event2 |2
2     |event1 |2
2     |event2 |1 
2     |event3 |4 
3     |event1 |3 
4     |event3 |5 
4     |event4 |5 

我怎么能得到这个表？(关于列事件{索引}计数总和)
仅使用能力 BigQuery(或 SQL)

userId|event1 |event2|event3|event4
----------------------------------
1     |1      |2     |0     |0    |
2     |2      |1     |4     |0    |
3     |0      |0     |0     |0    |
4     |0      |0     |5     |5    |

Answer 1

如果您只有以下几个事件适合您 - 您将需要构建尽可能多的相应行，因为您有不同的事件。如果预期事件的数量不变 - 您总是可以轻松地构建一次此类查询，然后使用它

SELECT 
  userID,
  SUM(CASE WHEN event = 'event1' THEN [count] ELSE 0 END) AS event1,
  SUM(CASE WHEN event = 'event2' THEN [count] ELSE 0 END) AS event2,
  SUM(CASE WHEN event = 'event3' THEN [count] ELSE 0 END) AS event3,
  SUM(CASE WHEN event = 'event4' THEN [count] ELSE 0 END) AS event4
FROM YourTable
GROUP BY userId

如果你需要更动态的东西 - 看看非常相似的例子 https://stackoverflow.com/a/36623258/5221944

在您的情况下，构建动态 sql 的查询将如下所示

SELECT 'SELECT userId, ' + 
   GROUP_CONCAT_UNQUOTED(
      'SUM(IF(event="'+event+'",[count],0)) as [d_'+REPLACE(event,'/','_')+']'
   ) 
   + ' FROM YourTable GROUP BY userId ORDER BY userId'
FROM (
  SELECT event FROM YourTable GROUP BY event ORDER BY event
)

请注意下面的行

      'SUM(IF(event="'+event+'",[count],0)) as [d_'+REPLACE(event,'/','_')+']'

它确保您的偶数名称符合字段/列名称的要求
如果您的事件总是看起来像 event1、event2 等，您可以简化此行并使用

      'SUM(IF(event = "' + event + '", [count], 0)) as ' + event