Increasing value of a variable (and subsequent values of the variable) by 1

[Janey Sewell]

Hi,

I think this will be relatively straight forward to anyone who has more experience with Stata than me - which is probably most of you on here.

I am trying to increase the value of a variable 'visit' by +1, but also require the subsequent values of 'visit' for that particular id to increase by +1 as well.

Here's my example data:

id visit dup datesubmitted
01 1 1 10dec2015
01 1 2 10apr2016
01 2 1 11aug2016
01 3 1 21dec2016
01 4 1 14apr2017
01 5 1 15aug2017

In the above example I am trying to get the second visit 1 in which dup == 2 and datesubmitted is 10apr2016 to take on the value of 2, and then all the subsequent visits to increase by 1.

i.e what i would like to happen is this:
.:
id visit dup datesubmitted
01 1 1 10dec2015
01 2 2 10apr2016
01 3 1 11aug2016
01 4 1 21dec2016
01 5 1 14apr2017
01 6 1 15aug2017


But when I use the command
by studyid: replace visit = visit[_n+1] if dup==2

this only increases 'visit' by 1 when dup==2 (as expected) but i can't work out the code to increase the subsequent visit numbers.

Any ideas would be really appreciated.

Thanks,

Janey

[Clyde Schechter]

If I'm understanding it correctly, you don't need that variable dup. You just want to renumber the visits in chronological order, right? If so:

by  studyid (datesubmitted), sort: replace visit = _n

[Janey Sewell]

Thanks, unfortunately it isn't a case of renumbering the visits in chronological order as i have other id's that have been allocated visit numbers that i don't want to be renumbered (and aren't necessarily in chronological order). The difference is that those id's (that i don't want renumbered) don't have dup value of 2, only 1.
I created the dup variable to count the observations in which visits were assigned the same visit number, now that they are identified i want to replace the repeated visit with n+1 and then all the subsequent visits with n+1.

Hope that makes sense.

[Nick Cox]

Janey: You can try Clyde's code safely by creating a new variable

egen ever2  = max(dup == 2), by(id)  
by  studyid (datesubmitted), sort: gen VISIT = _n if ever2

[Janey Sewell]

Great, thanks so much Clyde and Nick.