Calculating the difference between two paired observations ignoring missing values

[claire nicole]

Hi,

I'm sure there's a simple way to do this but I can't seem to find a solution.

I have two variables, one containing both the start and peak values of a signal (stored in start) and a second variable indicating where there's a pair (i.e. the start (location == 1) and the peak (location == 2) of each signal increase (sometimes there's no peak)).

I want to create a new variable to store the difference between each pair (i.e. the amplitude of the signal: peak – start for each pair and store this at location == 2 in the new variable). I'm not sure how to loop back over the missing values and unpaired 1s from the peak value at location == 2 to the start value at location == 1 for each pair. Can someone help please? :-S

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(start location)
3.7167106 1
        . .
3.9673915 2
        . .
        . .
        . .
 3.625683 1
        . .
 3.852861 2
        . .
        . .
 3.636759 1
        . .
        . .
 3.811554 2
        . .
        . .
        . .
        . .
        . .
        . .
        . .
 3.471214 1
        . .
        . .
        . .
        . .
        . .
        . .
        . .
        . .
 3.317838 1
        . .
        . .
        . .
        . .
        . .
  3.26407 1
        . .
 3.495139 2
        . .
  3.29939 1
        . .
        . .
        . .
        . .
        . .
        . .
        . .
        . .
        . .
        . .
 3.233487 1
        . .
        . .
        . .
        . .
        . .
 3.188093 1
        . .
        . .
        . .
        . .
        . .
        . .
        . .
        . .
        . .
        . .
        . .
3.0768125 1
        . .
        . .
        . .
        . .
3.2485664 2
        . .
        . .
        . .
        . .
        . .
        . .
 3.096682 1
        . .
        . .
        . .
        . .
        . .
        . .
        . .
        . .
        . .

[Nick Cox]

This may help:

gen locn = location
replace locn = location[_n-1] if missing(locn) 
gen prev = start[_n-1] 
replace prev = prev[_n-1] if missing(prev) 
gen locn_prev = locn[_n-1] 
gen change = start - prev if abs(locn - locn_prev) == 1 
. 
. list if change < . 

     +--------------------------------------------------------------+
     |    start   location   locn       prev   locn_p~v      change |
     |--------------------------------------------------------------|
  3. | 3.967391          2      2   3.716711          1    .2506809 |
  9. | 3.852861          2      2   3.625683          1    .2271779 |
 40. | 3.495139          2      2    3.26407          1    .2310688 |
 42. |  3.29939          1      1   3.495139          2   -.1957488 |
     +--------------------------------------------------------------+