I’ve really been enjoying SQLight.jl as a back end that just works. Is there any high level documentation on the relationship between SQLight.jl, Tables.jl, and DataFrames.jl?
Background: Unfortunately, the package I’m currently working on needs to also support text files as a data source for some users (in my area of scientific computing, databases are not widely used or understood), so DataFrames seems like the best representation since it can be populated by an SQLite response or a text file. However, I’m getting confused about the difference between Tables.jl and DataFrames.jl, since they seem to be very similar.
Tables.jl is an interface library, it defines generic functions, any type that implements those functions can be treated like a table source. This is true of DataFrames, but it’s also true of functions in the CSV library, the SQLite library, etc etc. It’s basically just ways to access data inside whatever types you are storing them.