csv yaml is a nice feature but most of its functionality can be handled by fread arguments like colClasses. What cannot be easily handled by fread args are attributes of data. Examples where it could be useful:
- we can store a factor column as integer, and in yaml metadata provide
factor class as well as a list of factor levels (that obviously won't be feasible for many levels)
- we can store POSIXct field as numeric and in yaml metadata provide
POSIXt class as well as origin and tz attributes.
Related to #1656
csv yaml is a nice feature but most of its functionality can be handled by
freadarguments likecolClasses. What cannot be easily handled byfreadargs are attributes of data. Examples where it could be useful:factorclass as well as a list of factor levels (that obviously won't be feasible for many levels)POSIXtclass as well asoriginandtzattributes.Related to #1656