TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Ask HN: How is a data warehouse different to an SQL RDMBS?

2 点作者 herodoturtle将近 2 年前
Hey folks, forgive the potentially dumb question.<p>I&#x27;ve been using standard RDBMSs (mostly MySQL and PostgreSQL) for a little over 20 years, but I&#x27;ve never dabbled in Data Warehouses. I&#x27;m trying to understand the technical differences between the two.<p>The top answer on (my) google search for &quot;differences between a database and a data warehouse&quot; is that &quot;a database is any collection of data organized for storage, accessibility, and retrieval. A data warehouse is a type of database the integrates copies of transaction data from disparate source systems and provisions them for analytical use.&quot;<p>That speaks more to the purpose of the Data Warehouse, and less to its underlying structure.<p>I&#x27;ve also come across articles that espouse the benefits of a Data Warehouse in terms of OLAP (versus OLTP). I understand the basic premise of this, but yeah I&#x27;m just trying to picture what it is about a Data Warehouse under the hood that makes it superior.<p>Can you kind folks please point me in the right direction? Technical depth welcome.<p>Thanks!

2 条评论

alex_lav将近 2 年前
IME a warehouse is usually accommodating of structured and unstructured data, has decoupled storage and compute, and is optimized for analytics&#x2F;big data (as opposed to app-tier databases which are normally optimized for throughput).<p>I think the buzzword is really centered around intent, not actual technical ability. You _could_ use Postgres as a data warehouse, given it supports JSON types, but you wouldn&#x27;t be taking advantage of all the innovation that&#x27;s gone into the warehousing space.
mrkeen将近 2 年前
I don&#x27;t think there&#x27;s an &quot;under the hood&quot; to speak of. A Data Warehouse could be built out of RDBMSs or something else.<p>I think it&#x27;s mostly a buzzword, but if I had to define it, I&#x27;d say that it&#x27;s a data store which isn&#x27;t directly involved in operating your product, rather, it&#x27;s data that you gathered and stored outside of the product, to allow data geeks to run their analysis jobs without them needing to touch your production database.