What inspired this wish list request?
Testing results very often include non-numeric entries such as "N/A" or "Nil" in columns consisting of otherwise continuous numeric data. This causes JMP to class it as "Character" data, resulting in the need to recode and then reclassify the column.
What is the improvement you would like to see?
I'd like JMP to be able to bring up an error message to say: "Column X's data is continuous and numeric apart from rows 35 ("N/A") and 37 ("Nil"). Would you like to change these to numeric entries?" "Yes" would then bring up a window where I could recode these two rows and it would automatically be classified as continuous numeric.
Why is this idea important?
It's easy to stick to rules for data recording if you're generating and doing the recording yourself, but if you're contracting out or getting help from others, these errors can sneak in (and I suspect they're pretty common!). This sorts of studies tend to have large data sets and to be able to work with the data without having to recode would be ideal.