For examples of atmospheric distortion of sunlight, a good search term is sunset mirage.
However, if the sun is high in the sky on a clear day, the atmospheric distortion is usually small compared to the size of any features of the sun. Most other stars, viewed in most optical systems, are true point sources of light, so any atmospheric distortion moves the entire image of the star.
This distinction between point sources and disk sources is sometimes given as an explanation of why stars twinkle and planets don’t. Unfortunately that’s a myth to begin with: the naked-eye planets will also twinkle, if the “seeing” is bad enough. Atmospheric distortion comes on a continuum, from twinkling stars to twinkling planets to green-flash sunsets.
There are three reasons why rays of sunlight and starlight are approximately, rather than absolutely, parallel. You are asking here about scattering, which would take parallel rays and distort them. But there is intrinsic angular dispersion in light rays as well. Light from a point source, like a distant star, is diverging, in that it goes in all directions. For an extended source like the sun, the light from any point on the source diverges, but there you also have light from many different parts of the Sun converging on any particular point. This finite-size effect is why the Moon’s “umbra” is cone-shaped, rather than cylindrical, and therefore why we sometimes get annular eclipses.
These two intrinsic angular dispersions are much larger for light from the Sun than for light from any other star. The biggest contribution to non-parallel sunlight is the finite-size effect, which gives a random ray in a bundle of sunlight a random angle of up to a half-degree. For starlight the intrinsic dispersion is very much smaller. When atmospheric scattering varies the direction of starlight by more than about about 1/60th of a degree, which is the resolving power of the human eye for point sources, then you start to see variations in the apparent position, size, and/or color of a star, and therefore say that it is twinkling.
There are many occasions when the half-degree dispersion in sunlight is just not important, so we approximate that “rays of sunlight are parallel.” Likewise there are many occasions where the divergence of starlight matters, in which case we talk about rays of starlight bring not-quite-parallel. But in general, sunlight is less parallel than starlight. Context matters.