Would it make sense to register CSV.jl with FileIO? about csv.jl HOT 5 CLOSED

That would be nice, but we'd have to choose why kind of object to return. A DataFrame?

from csv.jl.

Yes, it should return whatever CSV.read("mycsvfile.csv") would return. i.e. use the defaults for all other arguments.

from csv.jl.

Couldn't it just return a DataStreams.Sink type? If you wanted a DataFrame, you could use something like DataFrame(load("myfile.csv"))?

from csv.jl.

If someone wants to take this up, feel free, but I don't personally have a strong use-case.

from csv.jl.

Note also that you can read CSV, Feather, Excel, Stata, SPSS and SAS files via FileIO.jl today, I integrated all of that with IterableTables.jl. So DataFrame(load("filename.ext")) and load("filename.ext") l> DataFrame work for all those file extensions. Saving also works for CSV and Feather files.

from csv.jl.