Consider adding `col_raw`

[bart1]

I just encountered a data specification that contains hexadecimal coding (http://vocab.nerc.ac.uk/collection/MVB/current/MVB000377/). It contains strings like: 11e923067dfdf1485e802752. As R has a base type for raw data it might be worth adding a collector to read this data correctly. It would save a lot of memory to not store this as characters.

[bart1]

I don't seem to be able to add the labels feature and colspec