Instead of separately deriving the vertical force from a volume argument, it can be obtained directly by integrating the pressure around the wetted half of the cylinder, which also confirms the volume-based shortcut used in the other method.
Place the origin at the centre of the circle, and describe a point on the wetted (water-side) surface by the angle $\theta$ measured from the top point, so its depth below the free surface is $h(\theta) = R(1-\cos\theta)$ for $\theta$ from $0$ (top) to $\pi$ (bottom).
The pressure there is $p(\theta) = \rho g R(1-\cos\theta)$, and it acts radially toward the centre. Its vertical component, integrated over the whole wetted arc length $R\,d\theta$ and length $L$, works out to $$F_v = \rho g R^2 L\int_0^{\pi}(1-\cos\theta)\cos\theta\,d\theta$$
Evaluating the integral, $\int_0^\pi(1-\cos\theta)\cos\theta\,d\theta = -\pi/2$, and after accounting for the sign convention (the lower quadrant pushes up more than the upper quadrant pushes down), this reduces to the clean closed form $$F_v = \rho g \frac{\pi R^2}{2} L = 1000\times9.81\times\frac{3.14\times9}{2}\times10 = 1{,}386{,}153\ N$$
This is exactly the weight of a half-cylinder of water of the same radius and length, confirming the shortcut. Combining with the horizontal force $F_h = \rho g\cdot2R^2L = 1{,}765{,}800$ N found from the projected area: $$F = \sqrt{F_h^2+F_v^2} = \sqrt{(1{,}765{,}800)^2+(1{,}386{,}153)^2} = 2{,}244{,}876\ N$$
\[\boxed{F \approx 2.24\times10^6\ N}\]