stata - Reducing code when creating prior year test score in opposite subject -

i have data looks , variable i'm trying create math_score:

year     id    subject    score    math_score     2011      1       m          30         30 2011      1       r          40         30 2012      2       m          50         50 2012      2       r          60         50 

my colleague , have come following way of doing this:

bys id year: egen math = mode(score) if subject=="m" id year: egen math_score = max(math)          

or:

bys id year: gen math_score = score[_n-1] replace math_score = math_score[_n+1] if math_score==. 

ultimately going used in lagged variable denote prior year test score in subject , in opposite subject. we're sure there's more elegant way of doing less code can't seem think of one. ideas?

it's lot easier in one:

clear  input year     id    str1 subject    score     2011      1       m          30         2011      1       r          40         2012      2       m          50         2012      2       r          60         end  egen math_score = mean(score / (subject == "m")), by(id year)  

documented here.

we're dividing true-or-false expression evaluated 0 or 1. if result 1, score used is. if result 0, result missing , ignored, precisely want here.

that said, may better off different data structure:

 clear   input year     id    str1 subject    score         2011      1       m          30             2011      1       r          40             2012      2       m          50             2012      2       r          60          end   reshape wide score, i(id year) j(subject) string   list, sepby(id)        +-----------------------------+      | year   id   scorem   scorer |      |-----------------------------|   1. | 2011    1       30       40 |      |-----------------------------|   2. | 2012    2       50       60 |      +-----------------------------+