2个事实表之间的数据仓库关系具有不同的谷物，以便在报告工具中进行钻探？
我已经阅读了DWH工具包，并在堆栈溢出上进行了搜索。 有不同的商店，总部希望有一份每月收入的报告。 所有数据均在DWH中归一化，并为报告工具创建了周期性的事实表，以便总部可以看到每月收入。
半年后，请求更改。总部希望能够在报告中钻探，以查看每个商店的更多详细信息。
由于这是汇总的数据，因此不可能向下钻取。如何解决这个问题？
- 创建一个具有较低谷物的事实，并通过日期维度加入到下谷物的日期（DWH工具包中的p.81上的kimball show a日期维度），我认为可以用来从更高级别的谷物事实到较低的事实。加入谷物事实。
- 创建一个新的交易星模式，并在报告工具（例如PowerBi）中进行聚合。由于可以轻松完成较低的谷物聚集。
I've read through the DWH Toolkit and searched here on Stack Overflow.
There are different stores and HQ wants to have a report with their monthly revenue.
All data is normalized in the DWH and a periodic fact table is created for the reporting tool so that HQ can see the monthly revenue.
After half a year there is a request for change. HQ want to be able to drill down in the report to see more details for each store.
Since this is aggregated data a drill down is not possible. How to solve this?
- Create a fact with a lower grain and join via the DATE dimension to the lower grain (Kimball shows on p.81 in the DWH toolkit a date dimension) what I think is possbile to use to go from a higher level grain fact to a lower level grain fact with a join.
- Create a new transactional star schema and do the aggregation in the reporting tool (for example PowerBI). Since there is a lower grain aggregation can be easily done.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
您可以在下部谷物上创建一个事实表。最好的做法是始终以最低的谷物创建事实表 - 这样您就不会得到现在面临的确切问题。
您的BI工具可以进行聚合，或者，如果您面临性能问题，则可以创建自己的汇总事实表。从一个事实表到另一个事实表的钻探通常取决于您的BI工具
You create a fact table at the lower grain. It is best practice to always create fact tables at the lowest possible grain - so that you don't get the exact issue you are now facing.
Your BI tool can do the aggregation or, if you are facing performance issues, you can create your own aggregated fact tables. How you drill down from from one fact table to another is normally dependent on your BI tool