英文:
interactive cluster vs SQL warehouse to connect to Power BI
问题
I was reading about how to connect Power BI with Databricks and found this article from Microsoft and one of the requirements is:
> An Azure Databricks cluster or Databricks SQL warehouse
So, I would like to know which one is the best option in performance or other reason to choose one of them, also I´m not sure why in the article doesn't mention the serverless option to connect.
英文:
I was reading about how to connect Power BI with Databricks and found this article from Microsoft and one of the requirements is:
> An Azure Databricks cluster or Databricks SQL warehouse
So, I would like to know which one is the best option in performance or other reason to choose one of them, also I´m not sure why in the article doesn't mention the serverless option to connect.
答案1
得分: 1
通常情况下,SQL仓库有多个优点:
- 它针对BI工作负载进行了优化 - 使用经过充分测试的配置和实例类型以实现最佳性能,具有结果缓存等功能。
- 它默认启用了Photon - 在交互式集群上,您需要额外启用它,价格会增加。
- 通常成本较低(取决于使用的SKU),请参阅有关Photon的前一项。
- 您可以使用Serverless SQL Warehouse,它几乎可以立即启动,并且可以设置非常低的自动终止时间,因此在不使用时不会产生费用。
我只能想象在以下几种情况下交互式集群可能会有益处:
- 数据量非常小,因此可以使用非常小的实例来处理它。
- 交互式集群已经用于非SQL语言的开发,因此您可以直接使用它。
英文:
In general SQL Warehouse should be better for multiple reasons:
- It's optimized for BI workloads - uses well tested configurations & instance types to achieve best performance, has results cache, etc.
- It has Photon enabled by default - on interactive clusters you need to enable it separately with increased price
- Usually it costs less (depends on what SKU is used), see previous item about Photon
- You can use Serverless SQL Warehouse that starts almost instantly, and you can set very low auto-termination time, so you won't pay when it's not used.
I can imagine the benefit of the interactive cluster only in few cases:
- The amount of data is so small so you can process it with very small instance
- The interactive cluster is already used for development in non-SQL languages, so you can just use it
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库,让每个人都能够通过互相帮助和分享经验来进步。
评论