First off, let me start out with the disclaimer that the concept of a wall through which gravity cannot pass doesn't really mesh with any scientific description of gravity, so any answer you get is going to come with a large dose of interpretation.
That being said, the way I'm choosing to interpret this is analogous to the problem of what the electric field looks like if you put a metal plate with a hole in it in front of some electrostatic charge. You see, metal plates "block" electric charge-- more precisely, the electric field within one is always zero (alternatively, the voltage is always the same). The reason I'm talking about the electric field is that for charges that aren't moving, it works very similarly to how a gravitational field works.
In both cases you have some potential-- $V$ in the case of the electric field and $\phi$ in the case of the gravitational field-- whose rate of change tells you the strength and direction of the field. The only difference is that charge is the source of $V$, while mass is the source of $\phi$.
Now, if we have a planet on one side of the barrier and nothing on the other side, on the side with nothing, the gravitational potential will follow Laplace's equation:
$$\nabla^2 \phi = 0$$
Don't worry if you don't know what those symbols mean-- the important thing is that Laplace's equation has unique solutions. A fairly straightforward consequence of this is that if you have a solution to Laplace's equation for one set of boundary conditions, it can't be the same as the solution for different boundary conditions. But clearly, the setup with the gravity cancelling wall has different boundary conditions than those without it, since in the former every part of the gravity proof barrier has $\phi = 0$.
From this, we can definitively say that using the model you propose, the portal would affect gravity. As for the question of precisely how much, well, that depends on the shape of the portal and the mass distribution behind it. Even when specifying this, there likely isn't a closed form expression for any but the most simplistic limiting cases. To achieve any degree of accuracy, you would have to put in all this information to find boundary conditions at the barrier, and then use numerical methods to approximate a solution to Laplace's equation on the other side.
TL;DR There is a mathematical solution to what you propose, but it requires more information to calculate and a significant amount of effort.