There is no such control inside the Analysis dialog box. You can of course change the current sample and set it to be the whole dataset. In which case the analysis (and everything else in the preparation script) will be previewed on the whole dataset.
Note that the interface will hang of your dataset is too big (as a rule of thumb, compare to the default sample size, which is 30 000).
For datasets that fit in RAM, I would rather use the value_counts method of pandas.