Exploring Unknown Databases

[moxious]

Michael Hunger shared this link via twitter to his blog, addressing how to explore unknown databases.

Needs discussion; he's using javascript and a server API, and building complete statistics on various aspects of the database.

The discussion is about whether it is better to use a sampling approach (incomplete data, but something that scales really well) or to use complete statistics, which can be harder to scale quickly to large databases, but are more accurate.