Re: [Help-glpk] Ignoring useless data instead of returning error

[Jiri K]

Thank you!
Jiri


Andrew Makhorin wrote:
> 
>> Maybe I misunderstood your answer but I think you are refering to the
>> oposite problem than I have. Missing values for t[i] are not the issue,
>> trouble lies in values of t[i], which are in the csv data file (from
>> which I load the parameter values), but the coresponding i is not in the
>> set anymore.
> 
>> This is quite annoying, because I need to remove these values from the
>> csv when I disable the class and perhaps in 5 minutes add them back when
>> I reintroduce the class.
> 
> Oops, sorry.
> 
> You need to declare t to be indexed over a set, which is read from
> the database along with t[i] values, not over CLASS, assuming that
> CLASS is a subset of that set (for sanity check). For example:
> 
> param cl{i in ITEM} symbolic;
> set FOO;
> set CLASS:= setof{i in ITEM}cl[i], within FOO;
> param t {i in FOO};
> ...
> table tab_plant IN "CSV" "foo.csv" : FOO, t;
> ...
> 
> 
> 
> 
> _______________________________________________
> Help-glpk mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/help-glpk
> 
> 

-- 
View this message in context: 
http://www.nabble.com/Ignoring-useless-data-instead-of-returning-error-tp23540629p23559313.html
Sent from the Gnu - GLPK - Help mailing list archive at Nabble.com.