I'm studying probability theory, and as a result measure theory, and I've come to the derivation for the expectation of a random variable in terms of its density function, i.e $$\mathbb{E}[X]:=\int Xd\mathbf{P}=\int_{\mathbb{R}}xf(x)dx$$
I've managed to understand pretty much everything in the derivation with the help of this question's accepted answer, but the one thing I don't understand is the evaluation at x. What I mean is, I understand using change of variables and Radon-Nikodym to get $$\int Xd\mathbf{P}=\int_{\mathbb{R}} \mathbf{id}_\mathbb{R} f d\lambda$$ but I fail to understand what it means to evaluate the integrand at $x$ to achieve $xf(x)d\lambda(x)$. Do we just write it as such for notations sake, like how $dx$ is used for $d\lambda(x)$, or does it mean something different than the integral of the functions themselves?
For instance, what does $d\lambda(x)$ even mean? From my understanding (I'm reading Achim Klenke's Probability Theory), $d\mu$ is used to represent the integral being "with respect to $\mu$", but $\lambda(x)$ (which would be 0 anyway, right?) is now a number not a measure. Does it mean $(d\lambda)(x)$? If so, is $d\lambda$ some new measure, and it doesn't just serve as an indication of what measure is being used for the integral?