Determining multiple indexes of mapping equation that are conditional on a variable having a non-zero value

[KennethAnn]

Here I face a problem which seems as below, and I am curious about how to solve the problem.

When there are multiple indexes for mapping equations, we have use the code below to model them, for example,

@mapping(m, eq_var[i in Index1, j in Index2], var[i, j] - var_exg[i, j] ^ 2)
@complementarity(m, eq_var, var)

In some indexes, var_exg[i, j] = 0, which may cause many problems when the mapping equations are nonlinear, therefore, I want to treat the problem similar to Pyomo or GAMS, as the logic of below

for i in Index1, j in Index2
    if var_exg[i, j] == 0
        eq_var[i, j] = var[i, j]
    else
        eq_var[i, j] = var[i, j] - var_exg[i, j] ^ 2
    end
end

I have tried many methods, for example, I firstly found all indexes of non-zero value by using findall(x -> x !=0, var_exg), and used the indexes of non-zero value to map the equations, and for the indexes of zero value, the variables are kept to zero.

idx_non_zero = findall(x ->  x !=0, var_exg) # which defined a CartesianIndex
idx_non_zero_2 = [(k[1], k[2]) for k in idx_non_zero]

idx_zero = findall(x ->  x ==0, var_exg)
idx_zero_2 = [(k[1], k[2]) for k in idx_zero]

@mapping(m, eq_var_non_zero[(i, j) in idx_non_zero_2], var[i, j] - var_exg[i, j] ^ 2)
@complementarity(m, eq_var_non_zero, var[idx_non_zero])

@mapping(m, eq_var_zero[(i, j) in idx_zero_2 ], var[i, j] - 0)
@complementarity(m, eq_var_zero, var[idx_zero])

However, it failed and the possible reason is that mapping cannot recognize the multi-index [(i, j) in idx_non_zero_2], but only [i in Index1, j in Index2]. Is there any other possible solutions for this problem.

Thank you all!

[KennethAnn]

I found a right method, and I am not sure whether there are any much more convenient method?

@mapping(m, eq_var[i in Index1, j in Index2; var_exg[i, j] != 0], var[i, j] - var_exg[i, j])
for i in Index1, j in Index2
  (var_exg[i, j] != 0) ? @complementarity(m, eq_var[i, j], var[i, j]) : @complementarity(m, var[i, j], var[i, j])
end