Re: [Gretl-users] Gretl crash with stringify() and big dataset

On Tue, 18 Dec 2018, Allin Cottrell wrote: > On Tue, 18 Dec 2018, Artur T. wrote: > >> Just one question for clarification. Running the following yields an >> -- at least for me -- unexpected behaviour where series foo results >> in a constant ("Entry_c") even though S[i] is different for each i. >> >> <hansl> >> nulldata 14000 >> series foo = 1        # SET to 1 >> strings S = array($nobs) >> loop i=1..$nobs -q >>     S[$i] = sprintf("Entry_%d", $i) >> endloop >> stringify(foo,S) >> print S     # results in "Entry_1" for all obs >> >> series foo = 2        # SET to 2 >> strings S = array($nobs) >> loop i=1..$nobs -q >>     S[$i] = sprintf("Entry_%d", $i) >> endloop >> print S     # results in "Entry_2" for all obs >> </hansl> > > Shouldn't really be unexpected. > > Per stringify(), the strings from array @S are used only as needed to > encode the values of the series in question. Since the series you > define have only one value each (i.e. they are constant) they require > only one string, namely S[1]. Oops, not quite accurate: S[2] gets used in the second case. But that's in line with the documentation in the chapter on "String-valued series" in the User's Guide. Although a string-valued series in principle represents an encoding that starts at 1, we allow for the possibility that not all the integer values are included in a finite sample. Properly functioning example follows: <hansl> nulldata 20 series s = index strings S = array($nobs) loop i=1..$nobs -q   S[i] = sprintf("Value_%d", i) endloop stringify(s, S) print s --byobs </hansl> Allin _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users