Help with a basic loop

[Todd Greene]

Greetings Statlist users,

I could use some assistance with a seemingly simple loop.

As the merged datasets that I'm working with are panel, some values may be repeats. In this case I am essentially trying to loop through each unique patient (designated by "PAT_ENC_CSN") and tally the total number of days spent in 8D (designated by "days8D"). However, I also need to consider unique stays/time spent within 8D as some patients come and go through the unit during their overall stays (designated "stay").

So, essentially need a loop that totals days8D per unique "PAT_ENC_CSN" and unique "stay".

For the example patient below, total days in 8d would equal 15.

PAT_ENC_CSN	adm_date	dc_date	days8D	stay
2.51E+08	31-Jan-19	31-Jan-19	1	1
2.51E+08	8-Feb-19	18-Feb-19	11	2
2.51E+08	23-Feb-19	25-Feb-19	3	3
2.51E+08	23-Feb-19	25-Feb-19	3	3
2.51E+08	23-Feb-19	25-Feb-19	3	3
2.51E+08	23-Feb-19	25-Feb-19	3	3
2.51E+08	23-Feb-19	25-Feb-19	3	3
2.51E+08	23-Feb-19	25-Feb-19	3	3

Thanks,
Todd

[Wouter Wakker]

I'm pretty sure a loop is not necessary. If I understand you correctly this will do it:

egen stay_distinct = tag(PAT_ENC_CSN stay)
egen wanted = total(days8D * stay_distinct), by(PAT_ENC_CSN)

[Todd Greene]

Excellent. Thanks Wouter. I can then get the desired total sum by this: total wanted, over (stay_distinct) and taking the value where stay_distinct==1

Cheers!
Todd