Adding specific seed array per each call of sample or rnorm

[Melkiades]

@Melkiades I have been thinking about reproducibility as well and I think we will first have to implement a seed for each sample call as we did in chevron for this to be of use, since right now any changes/added variables in a dataset will change the values in any subsequent columns even if those columns are not altered (if I understand the seed issue correctly). If you think this would be useful you can open another issue for this. Let me know if you want me to work on any of these issues today as well! I do think these snapshots would be nice to have.

You could also implement the snapshots first for the purpose of reviewing the lubridate PR and then we can add in the seeds after if that’s easier for you.

Originally posted by @edelarua in #255 (comment)

ref: insightsengineering/tlg-catalog#25