Thank you all for helping me. It works now.   @Astounding @s_lassen @mkeintz  ... View more

Thank you both for helping me! I've tried both of your codes and the only I get to work is the last proposed by @s_lassen.   And it works! it only select years with >80% with the next year also >80%. BUT!   It do not select the FIRST year so in cases were individuals have many years according to the criteria it can look as this below:    ID Reirement_event 12278 2005 12278 2006 12294 2002 12294 2003 12294 2004 12317 2002 12317 2003 12317 2004 12317 2005 12317 2006     Can I add some "Select minimum year" for every ID into the program below?  This also happens in cases were people 2003 have 81%, then 79% in 2004 and then 82% in 2005 and 2006. Were there becomes a gap (but this is not a problem) ID Reirement_event 12322 2003 12322 2005 12322 2006     data LIB1.Test2; set LIB1.Test1; by ID; if percent>80 and lag(percent)>80 and not first.ID then do; retirement_event=Retirement_year-1; output; end; keep ID retirement_event ; run;     ... View more

Thank you s_lassen for your help. Very appreciated!!!   Since I am about to do survival analysis, I am interesting in duration (time to retirement) and event (retired y/n). So it is as you wrote in last sentence - I am only interested in one-life event.   I am using several datasets and merge them into one, currently in short-format-style (see attached illustrative Excel). Since my goal is to only have 1 variable with retirement year, for example. ID 4 become retired 2004, retirement_year=2004. Then I may drop the other information and only use the exact retirement date as in the Excel-example.   Since I only have one variable for retirement year, it doesnt matter wether I use long or short format? But maybe I have do make it LONG-format to be able to sort by year (percentage) and use the program you sent me?     Thanks!   If the Excel doesn't work: https://imgur.com/a/31naK ... View more

Thank you Astounding. Let me try and I will come back ASAP with more details if it works or not!   Thansk again! ... View more