Why not just code it yourself? data observations;   do ID=1 TO 100;     x1 = RAND('NORMAL',0,1);     x2 = RAND('NORMAL',0,1);     x3 = RAND('NORMAL',0,1);     output;   end; run; proc sql noprint ;   create table distance as     select distinct id,x1,x2,x3,sqrt((x1-y1)**2 + (x2-y2)**2 + (x3-y3)**2) as distance     from observations        , (select 0.5 as y1,0.5 as y2, 0.5 as y3 from observations(obs=1) )     order by distance desc   ; quit; data top10;   set distance (obs=10); run; proc print; run; ... View more

Are you sure you are transposing the same dataset? Is it possible that there is a format associate with the variable SEC that is only be found in one of the two SAS sessions? Look at the names of the variables when you do this: proc transpose data=sashelp.class out=x1;   id name ;   format name $5.; run; Look at the option DKROCOND.  %put %sysfunc(getoption(dkrocond)); You can set it to NOWARN and it will silently do nothing. You can set it to ERROR and it will cause and ERROR instead of just a WARNING. ... View more

You have many choices, and the right one probably depends more on both how you prefer to work and probably some details of what you are doing that I do not understand. If you are running Windows then you could install the SAS System Viewer 9.1, this is probably the sv.exe that your default settings is looking for.  You can still get it from the SAS download pages.  The last version was 9.1 and it cannot read version 9.2 64bit dataset created on Unix (Sun).  It can read 64bit created on PC and 32bit create on Unix.  I do not know about SAS 9.3 datasets.  This will open individual SAS datasets very nicely. You could get the "new" SAS Universal Viewer.  This should open all dataset formats, but it does not play as nicely as a browser tool for a single dataset.  It wants to open the whole library and if there are many other datasets there then you can waste a lot of time waiting for it to query each dataset for its metadata.  But perhaps it is more modern and can play nicely with your tools using some type of program to program communication protocol that I am too old to figure out. You could also just purchase a nice utility like STAT/Transfer that convert SAS datasets into dozens of other formats. I am a little worried about the path that you are trying to read the file from.   Are you trying to read a SAS dataset from the WORK library of a running SAS process? I do not think that is wise as it might crash SAS or your browser tool or both.  Do you have control over the process that you are running?  Can you have it write these files to another location?  If it is SAS code then just have it write the data out in a different format than a SAS dataset. SAS can write almost any format you want. I am also a little worried about your SAS work directory being under the /tmp partition.  That could be a recipe to crash your Linux server.  I would recommend that SAS work directories be in a different partition from the /tmp so that you do not run the risk of filling the /tmp directory and bringing your operating system to a halt. ... View more

If you table is small enough then why not just rotate it in Excel before reading it into SAS? http://office.microsoft.com/en-us/excel-help/rotate-data-by-converting-columns-to-rows-or-vice-versa-HP005203138.aspx ... View more

If that fixes your problem then most likely the underlying cause is that the macro variable has spaces in it. I usually use this simple code to fix that. %let bins_needed=&bins_needed; I find this usually happens when using PROC SQL to generate a count. proc sql noprint ;    select count(*) into :count from sashelp.class; quit; %put count="&count"; %let count=&count; %put count="&count"; 395  %put count="&count"; count="      19" 396  %let count=&count; 397  %put count="&count"; count="19" ... View more

There are probably ways of doing this better, but here is one way to read your example data into a hash that has unique records and use it to lookup the date for your transaction records. data dates;   input date ddmmyy10. number @@;   format date ddmmyy10.; cards; 10-10-2011 697 10-10-2011 122 09-10-2011 607 08-10-2011 607 07-10-2011 122 07-10-2011 607 run; data trans;   input number @@ ; cards; 529 122 697 607 888 run; data lookup ; if _n_=1 then do;   if 0 then set dates;   declare Hash H();   H.DefineKey ('number');   H.DefineData ('date');   H.DefineDone ();   /* complete hash table definition */   do until(eof);     set dates end=eof;     if h.find() then h.add();   end; end; set trans; if h.find() then date=.; put number= date=; run; number=529 date=. number=122 date=10/10/2011 number=697 date=10/10/2011 number=607 date=09/10/2011 number=888 date=. ... View more

Doesn't sound like you want to make a hash that has all of the values of created_date, just the latest date for each value of phone? Phone  Created_date 607 10-10-2011 122 10-10-2011 ... I am not that good with hash tables, but it seems like during the load of the hash from the source dataset you only want to add items when the key is new. ... View more

You are mixing up macro logic and data step logic. The reason that the %PUT is getting an error is that it is compiled BEFORE the data step runs.  So the CALL SYMPUT function has not been called yet. In a macro use the %LET statement to assign a value to a macro variable. %macro doit(m,n);   %do i=&m %to &n;     %let st&i=%eval(&i+1);   %end; %mend doit; %doit(1,2); In a data step use the DO statement to loop. %let m=1; %let n=2; data _null_;   do i=&m to &n;     call symputx(cats('ST',i),i+1);   end; run; PS: Do not use DO as the name of a macro as %DO is already a macro statement. ... View more

You can use PROC TRANSPOSE to rotate the grades into rows. Splitting into individual datasets is harder.  But we can query the original dataset and generate some code into macro variables to make it a little easier. Are you sure you need the individual files?  Why not just use the single rotated dataset and use a BY statement to process it by NAME? data have;   input name $ english science history math; cards; ram   65 48 89 49 Hari  79 69 90 98 gopal 87 35 30 90 ;;;; proc transpose data=have out=want(rename=(_name_=course col1=grade));   by name notsorted ; run; proc sql noprint ;   select 'grade_'||name        , 'if name='||quote(name)||' then output grade_'||name     into :dsnames separated by ' '        , :ifstatement separated by ';'     from have   ; quit; data &dsnames;   set want ;   &ifstatement ; run; ... View more

You need to add the MOD option to the FILE statement in the last data step. ... View more

Looks like you are just trying to generate T_QUICK based on which bin QUICK_RATIO falls into. Does not look like you even need to know how many bins there are to do that. Take a look at this example. Play with the condition in the UNTIL() clause to get it to pick the right bin. data v_quants;   input ratio @@; cards; 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 run; proc transpose out=quants(drop=_:) prefix=Qs; run; data Default;   input quick_ratio @@;   if _N_ eq 1 then set Quants;   array qs qs: ;   do i=1 to dim(qs) until (quick_ratio < qs(i) );     t_quick = qs(i);   end;   drop qs: i ; cards; 0.05 0.15 0.2 1.0 1.5 run; proc print; run;        quick_ Obs     ratio    t_quick 1      0.05       0.1 2      0.15       0.2 3      0.20       0.3 4      1.00       1.0 5      1.50       1.0 ... View more