inf values are definitly not missing-value indicators. They are no different than the value 7 or pi.
The solution in the 2nd link you provided is designed for matrices. It needs to be adapted to tables.
If your table only contains numeric scalars, you could convert the table to an array (matrix), apply that solution, and then convert back to a table.
If you want to apply that solution to specific columns of a table that contain numeric scalars, you can do so using
T.col(isinf(T.col)) = NaN;
where T is your table and col is the column name.
Note that you cannot remove numeric values from the table unless you remove the entire row. You can, however, replace their values with NaN or some other value.
If you run into any problems, show us what you've done, describe the variables, and the problem.