Thank you mohamed Here is my code (with the *1 technique) : data test ; set figures ; CA = A*1; run; In fact, in the log, it seems sas sees the data as e.g. "119." because it reads it like that (in the log) : 119€ But the € doesn't appear in my table ! I even ran a compress function before to suppress the € ! I have a way to do what I want (I substr a string as (length(myvariable)-1) and then *1) but it is not very good. So if you have a solution to handle that invisible € sign, it would be great ! Thanks ... View more

I would like to thank everyone here ! All of your answers have been very inspirational ! I applied the one provided by Astounding, and it worked just perfectly ! I have to say I have been quite amazed by that trick ! I called some of my colleagues to check on it and they were all "wooow". Do you have any links or resources where I could learn more tools like that ? Thanks a lot again, have a good day ! ... View more

Thanks Tom and RW9, I have asked my boss, and it can be both ! in fact, table B can be both following ways :      keep intermediate values With a structure like ID2 NEW_1 NEW_2 - ... - NEW_20 NEW_HIGH where each NEW_i reflects the successive changes in ID2 and, in the end, NEW_HIGH would be equal to NEW_20 (We need it to be called NEW_HIGH for our work) or      without intermediate values With a structure like ID2 NEW_HIGH where NEW_HIGH would be equivalent to the NEW_20 above I really don't see how to do that. I thought about merges but I can't figure out how to keep ID which doesn't change at all in that case. Do you see how to do that ? Both ways are ok (if you can, the simpliest one would be the best ahah) Thanks a lot ... View more

Yeah that's right ! In fact I got that data set prior in my study. Later we receive a list of "contracts" for which we need to put provisions (sorry my english is bad). But the problem is, I have no key to match the two lines since the amounts differ so we have to do that "manually". But yeah that's what I told my boss (that it was impossible) so now I am sure ! Thanks ... View more

Thank you RW9, I thought about something similar (much less well-said than the way you did ahah) but I gonna be honest, my boss is really stubborn, and he wants it in EG only. So you're telling me there is probably no way to do that ? ... View more

Yes that is incredible..And they do nothing to make the task easier. Above all, I need to automatize a whole process (beginning with the import of that xlsx). Well converting it to CSV make it not completely automatized but at least I can run the rest of my program... I did convert it to CSV but for ID I have the same trouble, it becomes digits with E+12 types. ... View more

Good morning Jaap, in fact the reason is simple : I receive that excel as an input from another department in my company, and I can only receive it that way. So, I absolutely need to import that excel file in SAS ... View more

I mixed all of your suggestions and it did exactly what I wanted ! Thanks a lot guys for your help ! Plus, it made me discover the arrays ! Great ! Have a nice day ! ... View more

Thank you both ! I am going to put a little example to show what I want to count (in bold the variables I already have, I want to create the variable N_MISS): VAR_1     VAR_2      VAR_3      VAR_4      VAR_5     SPAN     N_MISS      5               10           .               .                    10          5               2      2               3             .               .                    .               3               1      .               .               1               5                    .               4               2 ------------------------------------------------ I am sorry if I get it wrong but I have the impression that when writing things like nmiss(of var_1 - var_122) ; or                   n_miss=0; do I=1 to 122;      if var=. then n_miss=n_miss+1; end; I will get N_MISS values 2 3 3 instead of the 2 1 2 above My problem seems to me (again, I am probably wrong !) that my loop isn't fixed since for each observation (row) I don't use the same span to count. For my example above, 1st line I count on var1-var5, 2nd line I count on var1-var3 and 3rd line I count on var1-var4. Thank you very much for your help, those arrays are totally new to me (discovered them this morning ahah) ... View more

Thank you RW9, the problem is that your code seems to count the missing values for the 122 variables "VAR" for each observation. In my case I need for each obs to count the missing values from VAR_1 to VAR_j where j is the value of variable SPAN. ... View more