I am not aware that Proc Rank is capable of doing this. When it run into tied values, it ranks them the SAME regardless, because they are TIED. Those different options of tie= do not change this. Now in addition to Art's solution, you can also take advantage of _n_ to do the job: data have; input A       B     ; cards; 1222          180.00 1222          180.00 1222          260.00 1223          1 1223          4 1224          2 1224          5 1224          5 1224          5 ; proc sort data=have; by a b; run; data want; do _n_=1 by 1 until (last.a);     set have; by a;     Rank=_n_; output;   end; run; proc print;run; Regards, Haikuo ... View more

Another approach using mainly data step, though I can't guarantee more general application. data h1; input CID $   PID $     txtLabel :$10.; cards; 1-01      Root    level 1-02      1-01    level 2-00      Root    level0 2-01      2-00     level1 2-02      2-01     level2 2-03      2-02     level3 2-10      Root    level0_2 2-11      2-10     level1_2 2-12      2-11     level2_2 2-13      2-12     level3_2 ; data h2; input CID $ ; cards; 2-03 2-11 ; data want (drop=_:);    do until (last1);       set h2 (rename=cid=_cid) end=last1; do until (cid=_cid); set h1; if pid='Root' then _n+1; end; do until (cid=_cid); set h1; if pid='Root' then _m+1; if _m=_n then output; end; end; run; proc print;run; Regards, Haikuo ... View more

Add a few, ANYALPHA Function ANYDIGIT Function NOTALPHA Function NOTDIGIT Function And of course compress() plus lengthn(). Regards, Haikuo ... View more

Like this: data table1; input ID; cards; 2 3 5 ; data table2; input ID; cards; 1 2 3 4 5 6 ; data want1 want2; merge table2 (in=t2) table1(in=t1); by id; if t1 and t2 then output want1; if t2 and not t1 then output want2; run; proc print;run; Please note, you probably want to only keep the variables form table2 by using the data step options: data want1 (keep=) want2 (keep=); Edit: Art just has offered a great solution. Regards, Haikuo ... View more

Tian.Kong, Your fun is hurting my brain right now :smileysilly:. Haikuo ... View more

That is odd. I can't repeat your error. Please provide your error part of your log. Here is what I have just tried: data have;   input POINTS;   cards; 1.25 .77 .99 1.13 .03 1.03 .89 1.33 ; data want; set have; points=ifn(points>1.05,1.05,points); run; proc print;run; And this was my log: 587  data have; 588 589    input POINTS; 590 591    cards; NOTE: SAS went to a new line when INPUT statement reached past the e NOTE: The data set WORK.HAVE has 8 observations and 1 variables. NOTE: DATA statement used (Total process time):       real time           0.01 seconds       cpu time            0.00 seconds 601  ; 602 603  data want; 604 605  set have; 606 607  points=ifn(points>1.05,1.05,points); 608 609  run; NOTE: There were 8 observations read from the data set WORK.HAVE. NOTE: The data set WORK.WANT has 8 observations and 1 variables. NOTE: DATA statement used (Total process time):       real time           0.12 seconds       cpu time            0.01 seconds 610 611  proc print;run; NOTE: There were 8 observations read from the data set WORK.WANT. NOTE: PROCEDURE PRINT used (Total process time):       real time           0.00 seconds       cpu time            0.00 seconds And this is the output:    Obs    POINTS     1      1.05     2      0.77     3      0.99     4      1.05     5      0.03     6      1.03     7      0.89     8      1.05 Everything seems normal to me. Ifn() is a function, if is a statement. In this case, you can do it both ways. However, in the context of my code, it has to be ifn(). You need to rephrase the code to be able to use 'if'. For details, please refer to SAS manual. Haikuo ... View more

That is odd. I can't repeat your problem. You need to make sure that there is no empty rows in your 'have' table which may require you to expand  array range. Run this one to decide how many elements you are going to need in your array: proc sql; select max(ct) from (select (count(*)) as ct from have group by contract,fsdt) ; quit; HTH, Haikuo ... View more