First, there are times where you need multiple random numbers. Simply using multiple iterations of RAND() in a single statement won’t cut it, as you need to vary the seed. So I keep this snippet handy for when I need a bunch of random values in a single statement:
Click through for that as well as two more uses of
RANDOM(). This is my reminder that
RANDOM() generates data across a uniform distribution (every value in the range is equally likely to be chosen), making it great for these sorts of experiments but can look weird by itself if you’d expect non-uniform distributions of the data. For that, you would need some distributional trickery—though frankly, between the uniform and normal distributions, you’ve probably covered about 95-99% of test dataset needs.