Hi Fred, It would be useful to create for each issue separate threads. Otherwise one gets totally lost. Also, it makes it easier for others to search for specific things later. Thanks, Artur Am 31.05.2022 03:18 schrieb Fred Engst: > Hi, Sven, Artur, Jack and others, > Thanks for all the suggestions. > > 1) Helps on a more efficient script without loops: > Since gretl is not only a great research tool, but also a great > teaching tool, I’m trying to extent its usefulness. > For my online teaching of econometrics, I need to do an online exam by > creating individualized datasets with models for my students at > random. > What I’m trying to do here is to create two models, one with a full > set of variables, and the other after eliminated those seemingly > insignificant variables. > I will then ask the students to do a joint hypothesis test on that set > of variables, thus the "omit .... --auto” will not do. > Since the dataset was drown with a random subset of observations, I > can’t tell which set of variables will be insignificant before hand, > so I need a way to find them in a script. > I’m sure there are great one-liners out there, but I haven’t read > through all the documentations to figure it out, so I’m a bit lazy and > taking the easy way out by asking for help here. > > 2) A wish to save .gretl dataset from a scrip: > As you can see, the way I’m using gretl, I need to send my students a > .gretl dataset with all the models and perhaps graphs. As is, I need > to save each dataset manually, and it is a bit time consuming when > there are more than just a few students. I wish I can do this through > a script, perhaps using gretlcli to do that, if any of you can suggest > a way. > A year ago Jack suggested a method that seems a bit out of my reach. I > would like to see a sample code for combing gretl’s .gdt file with > models in a zip file become a .gretl file. > Thanks, > Fred > _______________________________________________ > Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it > To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it > Website: > https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/

On Mon, May 30, 2022 at 9:19 PM Fred Engst <engst.uibe(a)gmail.com&gt; wrote: You're talking about a gretl session file, not just a dataset (which is easily saved via script using either the GUI program or gretlcli, using the "store" command). Saving a session file via gretlcli is inherently impossible since that program has no notion of a "session", no named models or graphs. However, saving a session file via the command-line (script or console) in the GUI program, while it has not been possible before, is now implemented in git and snapshots. You use the "store" command, giving the name of the file to be saved the suffix ".gretl", as in store session01.gretl Note, though, that you cannot specify a list of series to save (as in standard "store" usage) -- or if you do, the effect is to save a gretl dataset (in the strict sense). with ".gdt" appended to whatever filename was given. There's no provision in session-saving to subset the existing series, though of course you can achieve that effect by deleting some series before doing the save. Here's a trivial example of the new usage: <hansl> open data4-1 m1 <- ols 1 0 2 g1 <- gnuplot price sqft store mysession.gretl </hansl> Allin