Does the date matter? IE your example chose the first one ... was that by design? EJ ... View more

The only way I have over come this is to put the email in a macro and send to the list one at a time. I believe I had to put some wait time in between the mailings so that mail servers did not mark them as spam. EJ ... View more

This is interesting Reeza ... if I understand you correctly you basically have a matrix that matches with the actually data that pinpoints the cells color. In my head you then parse that data (which variable is colored and what color it is) and create flags that you can merge back to the original data and use for the further requirements. Is this VBA code something you have used before? EJ ... View more

I run into this alot as well with not a lot of success being able to pull in and use the cell coloring in SAS. I usually just have to duplicate the excel algorithm that lead to the colors in SAS before even doing the next step (and this assumes there is logic that is repeatable behind the colorings). I would be interested to see others solutions for this as well! EJ ... View more

The max function in the datastep is not accross observations it is actually looking for max across variables so passing one variable does not make sense. Plus you are combining numerical results and character results ("0") in the same variable. It really isnt very clear what your objective is and why sql is out? If you just want to add the desc variable you could create the variable and merge it in. Or use macro variables to assign the max and min such as below: data a; input a1 b1 ; datalines; 1 4 2 5 3 3 4 2 5 9 ; run; proc means data=a min max noprint;   var b1 ;   output out=minmax (where=(_stat_ in ('MAX', "MIN")) keep=_stat_ b1); run; data _null_;   set minmax;   if _stat_ = "MAX" then call symput("max",b1);   if _stat_ = "MIN" then call symput("min",b1); run; %put &min &max; data b;   length desc $ 10;   set a;   if a1 = 1 then desc = compress("&max");   else if a1 = 2 then desc = compress("&min");   else desc = "0"; run; EJ ... View more

I dont think we have enough information to help answer those questions. If you have SAS available and the code was written for your specific situation, it isnt hard to run SAS code. The problem would be updating it if requirements changed. If this is just example code where conditions, table names, variables etc ... could all be different it might be easier to start from scratch with Oracle and just duplicate the requirements. You could probably find a SAS consultant that could outline each step in the program and describe what is going on so someone can duplicate that in Oracle. I couldnt even fathom a guess about how long that would take. EJ ... View more

Well I think converting 1200+ lines of code may be a little much to ask of the volunteer forum. Some the sql appears to already be written in places (commented out with *s) but some of the data steps will be hard to convert because of the processing that SAS enables through the data step. Probably going to have to find a SAS consultant to help rewrite if no one at your organization does SAS. EJ ... View more

If you switch to the advanced editor (upper right of the post box) the attach box will be at the bottom. EJ ... View more

Can you provide the code you trying to convert? Is it data step processes (steps that start with Data) or proc sql steps? SQL steps should translate pretty well ... data steps may or may not translate directly depending on what is occurring. Then if their are other procs in the code then it will really depend on which procedure it is and what is being done. EJ ... View more

As said without more information specific code will not be possible. Unless I am missing the issue here I dont think an array or do loop would be what I used. Do processing would be used however. Logic steps I would take: 1.I would sort by acct date 2.at the first of every acct I would set the counter to 1 and store the date value to date_tmp (retain this variable) 3.else I would check intck('month',date,date_tmp)=1 then do counter + 1;           else counter = 1; proc sort data=yourdata; by acct date; run; data newdata;      set yourdata;      by acct date;      retain date_tmp counter;      if first.acct then do;           counter = 1;           date_tmp = date;      end;           else do;                if intck('month',date,date_tmp)=1 then do;                      counter + 1;                                date_tmp = date;                end;                      else do;                          counter = 1;                           date_tmp = date;                       end;           end; run;           The syntax would depend a lot on the structure of your data, but perhaps the logic above can help you along. The code above isnt tested and may not be complete but perhaps provides a starting point. EJ ... View more

It has been awhile since Ive used VB but isnt there a shell function that allows dos command line like commands? Then you should be able to use regular batch like syntax to open sas and run a specific program. Batch syntax: "sas exe full path location" -sysin "sas program to run full path location" Add the -log and or -print to save out the log and lst to specific files. Eric ... View more

I dont think you can recursively create a variable with the lag function because the variable being created is going into the output dataset and not available to SAS PDV (I admit, I may have the SAS workings wrong here ... someone feel free to correct this). Why do you not want to use something like the following as suggested? data a;   set cake;   by type taste;   if first.type then mylabel=1;        else mylabel + 1; run; EJ ... View more