Add helper methods for DataFrames

[felipenoris]

All table methods avoid using DataFrames directly. I'm planning to add some helper methods to bring some sugar to the code. But this will add a depencency on DataFrames. Would like to know if someone is against this.

[YongHee-Kim]

Yes, I'm always using DataFrame, whenever I need to load something from Excel files, But I'm against adding a dependency on DataFrame. Because DataFrame is too widely used package, and It is quite likely that DataFrame would have major structural changes in the future for one reason or another.
I've used ExcelReaders before, and package author dropped DataFrame support after nullable & missing refactoring

Maybe creating a separate package or adding support for XLSX on ExcelFiles.jl might be a better way to do this.

[davidanthoff]

I’m very open to using this package here as the backend implementation for ExcelFiles.jl, which would solve the DataFrames issue. But, I’m also very nervous about taking on large dependencies, so I think I would only want to do that if the package here would essentially not add any dependencies over and above what it has right now.

Another option would be to move the code that is here into a package that might be called XLSXLib.jl or something, and have that package not take on new deps, and then XLSX.jl could depend on that and add some more integration with things like DataFrame or other packages. That would essentially allow folks like me to take a dep on XLSXLib, but not hold back nice features that you might like by taking on more deps.

In general I like a model where there is a low level package with minimal deps and other packages that provide more user friendly integrations. That is pretty much the model I use for most of the Queryverse file IO story, and it seems to work well.

[felipenoris]

Good to hear that! I guess I'll just leave this idea behind. Maybe I'll come back to this when Julia has native support for solutions like https://github.com/MikeInnes/Requires.jl .