It is important that you be able to replicate samples already produced, if required. For example, if fourteen items from 750,000 were selected, you need to be able to select the same fourteen items. For this reason, both cell and random sampling produce the same results given the same file and random seed.
If you set the random seed to 0, Analyzer will ignore the 0 and use a random number. Using a random seed (seed = 0) does not produce the same results, but since Analyzer reports the number it generates in the Command Log in response to a zero, even these samples can be reproduced by supplying a seed equal to the number Analyzer generated.