Sampling without replacement involves drawing samples from a dataset and simultaneously removing them from said set. This ensure you cannot draw the same sample twice1.

1

You can of course draw equal values, but not the exact same datapoint.