Hi,
working with panel data, I was puzzled by the reported time spans in
restricted samples. Consider the following example:
<hansl>
nulldata 2000
setobs 10 1:1 --stacked-time-series
genr time
genr unit
smpl time>1 --restrict
</hansl>
For me the first 'setobs' command responds with:
"1:01 - 200:10 (n = 2000)"
while at the end I get:
"1:1 - 200:9 (n = 1800)"
So I initially had t=1...10, then I restricted the time dimension to
t=2...10, but gretl tells me t=1...9.
Now, given the (relatively basic, if you'll excuse me) way gretl reports
panel samples this is probably not a bug, strictly speaking. But I think
it's a missing feature that creates confusion (at least for me).
thanks,
sven