@Dave-Allured I did some tests, a string attribute with multiple entries will be presented as an array of string by xarray in python. I don't think it has any concept of delimiters within the string itself (e.g. break on whites pace).
As for the actual topic of adding strings attributes to CF netcdf: are the CF version numbers meant to be semantic? (see https://semver.org/) if the answer is even close to "yes", then it would probably exclude adding the ability to have a string type represent any of the existing attributes currently defined in CF-1.x. Since all the more "complicated" values rely on some sort of character delimiter already, allowing them to exist in more than one data type is just added complexity without much benefit. -- You are receiving this because you commented. Reply to this email directly or view it on GitHub: https://github.com/cf-convention/cf-conventions/issues/141#issuecomment-410368695
