Originally posted by: ouaigooo

Hi,

Interestingly this code used to work with 12.6.3 but produces the following error with 12.7.1:

ASSERT FAILED: X& IloArray::operator[] (IloInt i) : Out of bounds operation: index superior to size of array    IBM_20171023_indexInEmptySet        Unknown    OPL Problem Marker

// model_file

tuple Tu_name_value {
  key string name;
  float value;
}

{Tu_name_value} DATA = ...;
{Tu_name_value} DATA_PERTURBATION = ...;

{string} N = ...;

float Qnow[v in N]   = item(DATA, <v>).value + ((<v> in DATA_PERTURBATION) ? item(DATA_PERTURBATION, <v>).value : 0);

execute { writeln("Qnow = ", Qnow); }

// data file

DATA = {
<"M_1", 1398.65718558>,
<"M_2", -8391.9431135>,
<"S", -20970.1664272>,
<"CH", -35667.2359386>
};

DATA_PERTURBATION = {};

N = {"M_1", "M_2", "S", "CH"};

What is wrong with it ? How come is it no longer supported ?

Thanks for your help !

Hi,

can you change

float Qnow[v in N]   = item(DATA, <v>).value + ((<v> in DATA_PERTURBATION) ? item(DATA_PERTURBATION, <v>).value : 0);

into

float Qnow1[v in N]   = item(DATA, <v>).value;
float Qnow2[v in N]   = ((<v> in DATA_PERTURBATION) ? item(DATA_PERTURBATION, <v>).value : 0);
float Qnow[v in N]   = Qnow1[v]+Qnow2[v];

?

regards

Originally posted by: ouaigooo

Hi Alex,

workaround works well ,thanks (but makes the code longer and less elegant)

I didn't make any syntax-error, right ?

So are you guys going to fix this bug in the next release ?

Kind regards -- Lucien