Hacker News Reader

I once read that the right way to use CouchDB is for every user to have its own database. However, how does this work with BI? Or with public data that should be known by all users? Do I create a single centralized DB just for that kind of data? Maybe aggregate data from all users' DBs? Genuinely curious.

View on HN

For public data, you can try to partition it in such a way that writes can be merged without any potential conflicts. E.g. a user's posts are in a separate partition.

I have never done this with CouchDB, but the technique is described in Martin Kleppman's __Designing Data Intensive Applications__.

You can replicate all per-user DBs into a central database today.

We are working on per-document-access-control at the moment, to support this use-case out of the box

Reply on news.ycombinator.com