On Tue, 6 Feb 2024, Sven Schreiber wrote: > > > The doc says that "delete" is only problematic in a loop, not in a > function per se -- but true enough, if I add the --force switch to > the delete command, then it does not complain anymore. So is this > just a missing piece of documentation? OTOH, deleting a locally > created series should not be a problem, should it? The error message is actually "You cannot delete series in this context". Deleting series is always something to be wary of, since it can invalidate existing lists, either directly or by causing other series to be renumbered. Since local series are automatically deleted when a function exits there's little reason to mess with deletion inside a function.

Am 07.02.2024 um 18:26 schrieb Sven Schreiber: Am 07.02.2024 um 17:45 schrieb Allin Cottrell: The error message is actually "You cannot delete series in this context". Deleting series is always something to be wary of, since it can invalidate existing lists, either directly or by causing other series to be renumbered. Since local series are automatically deleted when a function exits there's little reason to mess with deletion inside a function. The German message from which I re-translated uses the term for variable; this probably predates the time when I took over the translation, since I typically take care to use the word "Reihe(n)" for series. I'll check that. OK, so I looked at the respective po file (holding gretl's messages), and it actually seems that the original English version isn't always consistent in that respect. Here are some examples of message strings where perhaps "series" could be used instead of "variable": - ERROR: variable %d (%s), observation %d, expected numeric value - Variable %s cannot be renumbered - warning: missing value for variable %d, obs %d - variable %d (%s): non-numeric values = %d (%.2f percent) Should I change these occurrences of "variable" to "series" in the source?