DocumentDB查询需要意外的RU高 [英] DocumentDB Query Requires Unexpected High RUs

问题描述

我对存储在S3文档集合（最高的地区之一）的数据200MB。每个文件的大小大约为300KB。

I have about 200MB of data stored in a S3 document collection (the highest one). Each document is about 300KB in size.

但是，当我跑我的查询，我很惊讶地看到，它需要的RU 7245.81。由于S3是在2500个RU /秒，这个性能不会扩展我的应用程序。

But when I ran my query, I am surprised to see that it requires 7245.81 RUs. Because S3 is at 2500 RUs/Second, this performance won't scale for my application.

我只想问，如果我做错什么，如果有什么我可以做些什么来改善呢？

I just want to ask if I am doing anything wrong and if there is anything that I can do to improve it?

我的查询是这样的：

SELECT item.Id,item.Priority, va.something, wa.something... 11 fields in total
FROM model.Item item 
JOIN va in item.Content.Children 
JOIN wa in va.Children 
WHERE item.State != 5

我使用范围作为索引类型。但愿这种影响查​​询的性能？

I am using "Range" as the index type. Would that impact the performance of this query?

推荐答案

使用查询！=将需要扫描的，因为所有的索引条目都必须抬起头来，以消除不等于5.这将导致该值高个RU。只要有可能，请尝试重新编写查询作为平等或范围查询。

Queries using != will require a scan, since all index entries have to be looked up to eliminate the values that are not equal to 5. This will lead to a high RUs. Whenever possible, please try rewriting the query as an equality or a range query.

1. 如果item.State的可能值的数量是有限的，例如，1-10，那么你可以使用枚举有效的值，也就是说，item.State IN（1,2,3,4,6,7 ，8,9,10）


2. 如果这是不可能的，请在使用例如范围，item.State> 5或item.State＆LT; 5像拉里建议之上。

希望这有助于。请在微软科网arramac如果你想更详细地讨论给我发电子邮件。

Hope this helps. Please email me at arramac at microsoft dot com in case you'd like to discuss in more detail.

这篇关于DocumentDB查询需要意外的RU高的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！