Line by line calculation with lag

[PeterClemmensen]

So your data looks like this.. Rather than us correcting your code a bit, let me ask this.

What does your desired result look like, given this data?

data WORK.VARIATION6;
  infile datalines dsd truncover;
  input NO_POLICE:$20. DATE:DDMMYY10. MT_EA:32. CD_EVT:$8. MT_EVT:32. MT_EA_AVMVT:32. DeltaEAs:32. EAs:32. count:32. boucle:32. EAs_avant:32.;
  format DATE DDMMYYS10.;
  label NO_POLICE="NO_POLICE" MT_EA="MT_EA" CD_EVT="CD_EVT" MT_EVT="MT_BRUT";
datalines4;
V55900010,31/12/2019,147171.89315,,,,0,147171.89315,1,7,
V55900010,15/01/2020,,RDES,-1800,149675.17929,-1769.895376,145371.89315,2,7,147171.89315
V55900010,22/01/2020,,VINV,100000,148782.16658,100000,245371.89315,3,7,145371.89315
V55900010,15/04/2020,,RDES,-1800,227609.96392,-1940.46605,243571.89315,4,7,245371.89315
V55900010,15/07/2020,,RDES,-1800,237720.68701,0,0,5,7,243571.89315
V55900010,15/10/2020,,RDES,-1800,232272.14554,0,0,6,7,0
V55900010,31/12/2020,248976.1154,,,,0,0,7,7,0
;;;;

[SASdevAnneMarie]

Hello,

I added the data with results into my first message.
Thank you very much !

[Kurt_Bremser]

Please supply example data for variation5, and the variation6 you expect from this.

[SASdevAnneMarie]

Hello,
Thank you,
I added the data with results into my first message.
Thank you !

[SASdevAnneMarie]

I added the Variation5.
Thank you, Kurt.

[SASdevAnneMarie]

Finally I found the solution 🙂 Thank you !