Inspecting row variables for desired increasing trend (and detecting non-desired level or decreasing trend)

[Michael McCulloch]

In my dataset, a calculated field for age is being used to track subjects' progress. Code for a simulation is provided below.
Rather than manually inspect >1200 records, I wish to detect whether age (when observations are sorted on id & date)
increases (as in id==1), or does not change (as in id==2), or decreases (as in id==3). Is there a way to do this?

set more off
clear
set obs 9
input id age str9 date
    1    40.0    21jan2016
    1    40.5    15jun2016
    1    40.9 10dec2016
    2    40.0    21jan2016
    2    40.0    15jun2016
    2    40.0    21jan2016
    3    40.9 21jan2016 
    3    40.5    15jun2016
    3    40.0    10dec2016 end
gen sif_date = date(date, "DMY")
format sif_date %td
drop date
sort id sif_date
l, noo sepby(id)

[Andrew Castro]

I'm not sure how an age would decrease/stay the same (as the date increases), so I just want to check. Do you mean decreases/increases/stays the same across all dates or just date to date?

[Mike Lacy]

I presume you're trying to catch possible errors, so what about the following?

egen any_increase = total((age > age[_n-1])& (_n > 1) ), by(id)
egen any_decrease = total((age < age[_n-1]) & (_n > 1)), by(id)
by id: gen all_equal = (any_increase == 0) & (any_decrease == 0)

[Michael McCulloch]

Thanks Mike for the elegant suggestion, which did flag some errors. Much appreciated!
And apologies to Andrew for my somewhat ambiguous question.