rstan MCMC: Different squence of data resulting in different results, why?

Question

I am new to Stan and rstan.

I recently may find a weird issue when I worked on Markov chain Monte Carlo (MCMC). In short, for ex