Am 03.01.2019 um 13:35 schrieb Schaff, Frederik:
For example, with R you get the information right away. R imports the
data in its own data field type. I agree, it seems to be a feature
request then. As it seems to be a special request, I would not open a
ticket for it, if no one else sees a need for this feature.
There's no harm in opening a ticket, as this is not a guarantee it will
be implemented. It just serves to be reminded of it.
The thing is, gretl already reads this information alright, so it
wouldn't be a whole new thing. However, the question is how to store
that information, because essentially it amounts to having different
categories of missings, which gretl doesn't have. String-valued series
are not the answer, because the majority of values in those imported
series are numeric and must be interpreted as such.
In principle one needs some nested bundles I guess (not an array of
bundles though).
Also you have to watch out that the codes like -5, -1 etc. in your
series are not to be treated as regular numbers in the econometrics.
Some conceptual thinking needed I guess.
cheers,
sven