SuperCSV is great, but if you have hundreds of different csv files it's not very convenient to write Java parser for each and every one of them.
Instead it would be really beneficial if I could specify the cell processors in the header of the csv file. Like, for instance:
code[id,mandatory],name[optional],isFeatured[bool,mandatory],creationDate[date='yyyy/mm/dd hh:MM:ss'],price[double, optional]
my-code,My funky name,TRUE,2012/07/12 12:12:12,23.22
Well you get the idea. The [] brackets in the header contain the cell processors that need to be chained for each column.
I have actually built that myself by getting the array of headers and then iterate over each header column and split it and add the required cell processors. I think, though, that if that feature was out of the box it would help a lot of people.
Sounds great that you have built it already. Please share with us if you want it as a feature.