How do i scramble the subject id across the datasets without making any change to actual subject id?

[KemarRoach]

I have 4 datasets - A,B,C,D. Each dataset has around 500 observations. The subjects need to be scrambled within the available subject ID's. For example Actual subjects are 101,102,103,104 , they need to be scrambled as 103,101,104,102 . I am instructed to merge 4 datasets and apply scrambling for those which have subjects in dataset A. For example if dataset A has 101 as subject id and any of the datasets B,C or D has 101 as Subject id it should be scrambled with same number. Unmatched subjects should remain as it is .

Eg 1. A-101 B-101 C-101 D-101 -Actual subjects.

         A-103  B-103 C-103 D-103 - Scrambled subjects should be same across the datasets for a  particular subject.

Eg 2. A-104 B-105 C-109 D-104 - Actual subjects

         A-107 B-108 C-106 D-107 - Please observe scrambling for Dataset A and D.

Please help me with this requirement .

[Kurt_Bremser]

So you want consistent anonymization across a number of datasets?

[Reeza]

Are all 4 files guaranteed to have the same set of IDs or will there be IDs in some data sets that are not in others? What do you mean by "not change to actual subject id"? Can you show that with a before and after data set please?

Here's an example on creating an entirely randomized IDs:
https://gist.github.com/statgeek/fd94b0b6e78815430c1340e8c19f8644

[sbxkoenk]

Agreeing with @Reeza .

I think this is easy to do, but the question is not 100% clear to me neither.

Thanks,

Koen