K-S is only good for small sample size. Most of time, it is too sensitive for large sample size. If your QQ and PP graphs are OK, your data should be fine for normal distribution.
在stata里help swilk得来的资料:
swilk performs the Shapiro-Wilk W test for normality, and sfrancia performs the Shapiro-Francia W' test for normality.
swilk can be used with 4<=n<=2,000 observations, and sfrancia can be used with 5<=n<=5,000 observations;
sktest for a test allowing more observations.
也就是说样本量超过了swilk的范围时,最好用sktest