适用于小型数学数据但速度快且具有聚合函数的最佳数据存储解决方案
我正在为具有以下要求的项目寻找数据存储解决方案：
- 应用程序动态创建 商店里的容器/桌子。
- 在一小段时间内（两 例如几周） 桌子/容器数量巨大 并行插入。最后阅读 必须立即可用。
- 插入的数据很小，4 至 6 个数字列。
- 小查询/过滤支持是 必需，但不需要连接或交叉 查询。
- 需要执行一些聚合 函数如“计数”、“求和”、 “最大值”、“最小值”和“平均值”。
基本上，我需要类似 Windows Azure 表存储的东西，但具有聚合函数。
你会推荐什么？
I'm looking for a data storage solution for a project with these requirements:
- The application creates dynamically
a containter/table in the store. - For a small period of time (two
weeks for example) that
table/container gets a huge amount
of inserts in parallel. Last read
have to be immediately available. - The data inserted is very small, 4
to 6 numeric columns. - Small query/filtering support is
required, but no joins or cross
queries. - Is needed to execute some aggregate
functions like "Count", "Sum",
"Max", "Min" and "Avg".
Basically, I need something like Windows Azure Table Storage but with aggregate functions.
What would you recommend?
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（6）
RavenDB 支持您提到的所有内容以及更多内容。它的扩展性非常好，并通过 Map/Reduce 查询和 Linq 查询支持聚合函数。它也可以在内存中运行。
RavenDB supports all you mentioned and more. Its scales very well, and supports aggregate functions via Map/Reduce queries and Linq queries. It also can run in-memory.
MongoDB 也是一个不错的选择。它支持单节点安装的组（聚合）查询。如果您需要扩展，您可以创建分片集群并使用 Map/Reduce 来满足聚合需求 - 但与其他解决方案相比，Mongo Map/Reduce 性能并不是顶级的。
如果你需要map/reduce的高性能，你可以看看Hadoop
MongoDB is also a good choice. It supports group (aggregate) queries on single-node installation. If you need to scale you can create sharded cluster and use map/reduce for aggregation needs - but Mongo map/reduce performance isn't top level when comapring to other solutions.
If you need big performance in map/reduce you can take look at Hadoop
为什么不使用SQL Azure？它支持聚合功能并满足您的所有需求，看看此表
Why don't you use SQL Azure? It support Aggregate function and meet all of your needs, take a look at this table
如果不需要免费，kdb+ 将是一个不错的选择。它是一个主要用于金融领域的时间序列分析的系统。它支持您要求的一切，并且速度快得令人难以置信。有一个免费试用版，但要在生产中使用它......昂贵。它可以使用类似 sql 的语法进行完全查询，并且可以使用 apl 系列中名为 q 的专有语言进行完全编程。
http://www.kx.com/
If it does not have to be free, kdb+ would be a good choice. It is a system mainly uses in finance for time series analysis. It supports everything you ask for and is incredibly fast. There is a free trial version available, but to use it in production it is...expensive. It is fully queryable with a sql-like syntax and fully programmable with a proprietary language called q which is in the apl family.
http://www.kx.com/
对于问题正文中列出的要求，市场上任何有能力的 SQL 服务器都可以处理这个问题，尽管您需要设置一些索引以使事情变得超级快。
但这就像说，如果你想让你的车开得快，你必须把它放在D档并踩下油门踏板，所以就是这样。
For the requirements listed in the question body, any competent SQL server on the market would handle this, altho you would need to set up some indexes to get things to be super fast.
But that's like saying, if you want your car to go fast, you have to put it in D and press the gas pedal, so that's that fwiw.
确实需要更多有关要求和预算的信息……但是，由于人们只是提出一些答案，所以我将把这个问题纳入其中：
有一些非常酷且非常昂贵的硬件解决方案是为大规模并行运行而构建的SQL 查询，例如 Netezza Skimmer 或 TwinFin
http://www.netezza.com/data-warehouse-appliance-products/skimmer.aspx
几乎让我希望我们的数据仓库太慢 - 只是为了我能得到其中之一小狗。
Really need more info on the requirements and budget... but, since people are just throwing up some answers, I'll throw this one into the mix:
There are some very cool, and very expensive, hardware solutions built to run massively parallel SQL queries, such as the Netezza Skimmer or TwinFin
http://www.netezza.com/data-warehouse-appliance-products/skimmer.aspx
Almost makes me wish our data warehouse was too slow - just so I could get one of these puppies.