cumulative numbering within subgroup

[Tessa Jones]

Hello,

I have data wherein patients have multiple assessment types. I would like to create a new variable that numbers (1,2,3 etc) for the subgroup of Assessment.
I've tried the bysort command that i've seen in previous posts: bysort Assessment id: gen newvar=(_n)

but this does not give me the desired outcome.

This is what I'm trying to achieve:

id. assessment. newvar
1. SOC 1
1. dischargetype1. 2
1 dischargetype2. 3
2. SOC 1
2. dischargetype3. 2
3. dischargetype1. 1
3. SOC 2
3. dischargetype2. 3
4. SOC 1

Thank you!

[Øyvind Snilsberg]

bysort id: gen wanted = _n

[Nick Cox]

While I agree with #2 as quite possibly what you want, is there a time variable that you can make explicit? Otherwise put, what is the rule that allows you to identify 1, 2, 3 and so on?

[Tessa Jones]

Thank you! Yes, the time variable is ‘assessment date’.

[Nick Cox]

Good; then safer and more explicit code is more like

bysort id (assessment_date) : gen wanted = _n

[Tessa Jones]

This worked perfectly, thank you very much!