Hi, as (correctly) explained in the docs, in a panel context $pd gives the time-series length (=dimension). But isn't that a misnomer, or is it just me who's confused? It would be useful to get the standard periodicity information in panels as well. (Provided it's properly set with setobs.) Think about quarterly or monthly panels of macro data. I guess it's a historical gretl thing, and if the issue is backward compatibility, a new accessor like $ppd could serve that purpose. And something like $pdims could return a 2-element vector with the current cross-section and time dimension lengths. (The latter would be subject to subsampling and could be complicated by the fact that some units might be left out, so it wouldn't be always equal to max($unit)-min($unit)+1. Do you want me to provide a prototype in hansl?) thanks, sven