Change name of rows and columns in highspy

[FabianHofmann]

I am wondering whether there is the possibility to change the names of the rows and columns through the highspy interface.

Minimal Example

import highspy
import scipy.sparse

h = highspy.Highs()
h.addVars(2, [0, -10], [20, 0])
h.changeColsCost(2, [0, 1], [5, 4])
A = scipy.sparse.csr_array([[1, -3], [-2, 1]])
num_cons = A.shape[0]
lower = [-10, -20]
upper = [10, 10]
h.addRows(num_cons, lower, upper, A.nnz, A.indptr, A.indices, A.data)
h.writeModel('example.mps')

The program then warns:

WARNING: There are empty or excessively-long column names: using constructed names with prefix "c"
WARNING: There are empty or excessively-long row names: using constructed names with prefix "r"

There is the possibility to change the names in the LP model like

lp = h.getLp()
lp.row_names_ = ["c1", "c2"]
lp.col_names_ = ["x1", "x2"]

but that does not change anything when writing to file afterwards. Could you help me out here?

[jajhall]

Changing the names in a model is not currently possible, even with the C++ class. The only way to define names in a model that's not read from a file, is to create an instance of the HighsLp class, and populate its names.

In your case, as you're defining the LP by adding variables and constraints, once you've done this the route would be

lp = h.getLp()

populate its names, and then call

h.passModel(lp)

before

h.writeModel('example.mps')

I'll turn this into a discussion, as it's not a bug

[FabianHofmann]

Thanks @jajhall that makes sense.