How to design a database to save data whose structure is not always known during design, and could change later? -

Allows sites like IBM's Eyes, Swivel etc. stores and their users to visualize them. How do they design their tables? For example, if you save data from the data.gov site in the database and allow your users to work on it, how do you know about designing tables? To maintain any type of data the structure should be normal. For example, there are many data in data.gov, some of them are more complex than others.

The easiest way to answer your question is to use string-string dictionary. A popular structure and flexibility in its community is built with Python and Lua. You can be a domain expert for this by adding dimension of time - like, the hypertext does this.

Any data model can be sorted in string string dictionary I do not know the features, but MyScale has BDB back-end BDB's Core Data Structures string string.

ps, I'm half of a relational zealot, so if the data is important then I relationally model it: p