Right now the full actual question text appears in column names. These are almost always illegal variable names. But this is very useful info. How to balance?
Best I can think of right now is to offer an additional optional function that would set this full question text as a label attribute (I already have this implemented elsewhere) and truncate the column names to either a cleaned up version (janitor::clean_names() + truncation from the center), or to names more like what SurveyMonkey offers in SPSS, like q01_01.
The latter would aligned with what SM does for SPSS ... but the former is more obvious what they are. Maybe options for both?
Right now the full actual question text appears in column names. These are almost always illegal variable names. But this is very useful info. How to balance?
Best I can think of right now is to offer an additional optional function that would set this full question text as a label attribute (I already have this implemented elsewhere) and truncate the column names to either a cleaned up version (
janitor::clean_names()+ truncation from the center), or to names more like what SurveyMonkey offers in SPSS, likeq01_01.The latter would aligned with what SM does for SPSS ... but the former is more obvious what they are. Maybe options for both?