Perform additional condition after other conditions are met

[Yifei Liu]

Dear Statalist,

I am trying to generate a new variable NewCost based on another variable Cost. I want the generate the variable using data that meet three conditions (Time1<Time2, Type!=1, Time1 is the earliest). The last condition should be applied additionally among the data that are not exclude after the first two conditions. However, I don't want to drop any data because I still need the entire data table for next step of analysis.

cost	time1	time2	type	newcost
2.4	1997	1998	0	4.1
5.6	2002	1998	0	4.1
5.8	2001	1998	1	4.1
4.1	1996	1998	0	4.1
7.3	1996	1998	1	4.1

I tried the following code but couldn't get what I wanted.

egen newcost = mean(cost) if (time1<time2&type!=1)&time1==time1[1]

Any suggestion will be appreciated. Thanks!

[William Lisowski]

The following code seems to do what you seek.

* Example generated by -dataex-. For more info, type help dataex
clear
input float cost int(time1 time2) byte type float newcost
2.4 1997 1998 0 4.1
5.6 2002 1998 0 4.1
5.8 2001 1998 1 4.1
4.1 1996 1998 0 4.1
7.3 1996 1998 1 4.1
end

generate seq = _n

generate ignore = ! (time1<time2 & type!=1)
sort ignore time1
generate wanted = cost[1]

sort seq
drop seq

. list, clean

       cost   time1   time2   type   newcost   ignore   wanted  
  1.    2.4    1997    1998      0       4.1        0      4.1  
  2.    5.6    2002    1998      0       4.1        1      4.1  
  3.    5.8    2001    1998      1       4.1        1      4.1  
  4.    4.1    1996    1998      0       4.1        0      4.1  
  5.    7.3    1996    1998      1       4.1        1      4.1