Hi. I'm a big advocate of the SAS Certifications. My suggestion (before paying any more money for courses, guides or failed exams): Each Certification recommends specific training courses (and if training is your thing and you want to pay for it, then excellent. I'm also a SAS trainer, by the way.) Base Certification recommends knowledge equivalent to the Prog1 & Prog2 training courses. However, training is not a pre-requisite; you could learn your own way. So look at the agendas for each training course; some of it you will know already, other stuff you will have to go & learn and/or practice (or do a training course). But at least it serves as a list of what you need to know in order to pass certification. Good luck! ... View more

Hi (suicidedoor?) - do you specifically have to use a JOIN clause. Your desired result could be achieved in SQL using SET operators. ... View more

Moataz, Sandhu - presumably you obtained the installation media from your local SAS offices. What does their tech support suggest? ... View more

I'd also recommend checking out of 's excellent blog of all this SAS Admin: http://platformadmin.com/blogs/paul/ ... View more

In both cases, I don't see why Unix commands are needed, everything should be able to be done from within SAS. 1. PROC EXPORT won't help here; the simplicity of most PROC STEPS means some specific requests (like this) can't be done that way. Alternative, using DATA STEP, which also does not require Unix commands, plus it means you are not unneccesarily reprocessing either the source data set or the target DAT file: filename MyOut "./xx.dat"; data _null_;   file MyOut; /* Insert options of choice here */   set MyInputTable nobs=RecordCount end=IsLast;   if _N_ eq 1 then do;     /* All your header stuff goes here */     put "Records: " RecordCount;   end;   /* Output each line, apply formats as desired */   put _ALL_;   if IsLast then do;     /* All your footer stuff goes here (if any) */   end; run; filename MyOut clear; 2. Similar DATA STEP methodology applies here: filename MyInFile "./yy.csv"; data MyOutput;   infile MyInFile dsd end=IsLast; /* Insert other options of choice here */   if _N_ eq 1 then do;     input RecordCount; /* Syntax of input statement depends on the layout of text file */     put "Records: " RecordCount;   end; /*   Accrue an incremental count for every records read - ActualCount.   NOTE: I don't know if your first line also contains data, therefore   do not know if the next line should or shouldn't be within an "ELSE"   associated with the above "IF". Plus this may affect whether you are   comparing ActualCount with RecordCount, or ActualCount with (RecordCount-1) */   input Var1 Var2 Var3; /* Syntax of input statement depends on the layout of text file */   ActualCount+1;   if IsLast then do;     if RecordCount=ActualCount then put "Good!";     else do;       put "ERROR: Count mismatch " RecordCount= ActualCount=;       abort; /* Abort, etc - insert preferred action here */     end;   end; run; filename MyInFile clear; ... View more

Could you please share your final result? ... View more

Agree with - sounds like your "AB" variable is character, in which case your data step will need to convert it to a numeric using the input() function. data MyOutputData; set MyInputData; /* Which contains the (presumably character) variable AB */ format ab  BEST15.; /* Your choice of format is actually irrelevant for this exercise */   ab_num = input(ab,best15.); /* Note: The INFORMAT you choose will depend on the character "layout" - I've guessed BEST. here); drop ab; /* Optional */ run; proc summary..... etc. If this does not help, please post the results of a PROC CONTENTS, plus the first few rows of a PROC PRINT of the AB variable. ... View more

- Is it a "Macro community" question? ... View more

It depends on how you compiled/stored the macro. "session compiled macro" is not specific enough for us to tell how you compiled it: Did you just compile the macro (without the /store option) to the WORK.SASMACR catalog? If so, you might be in trouble. Did you store it in a non-WORK catalog, using the MSTORED option, and %macro xyz / store option? If so you might be in luck. Another thought - why did the Enhanced Editor close suddenly? If SAS crashed, it may not have cleared your WORK folder, so you might at least be able to retrieve the sasmacr.sasb7cat file and stick it somewhere where you could re-invoke the macro and examine the log (as explained in 's reply. Other suggestions to avoid future repetition: Set your SAS startup options to have an ALTLOG keep a copy of everything you generate. (But remember to purge it periodically..) Keep your macros as *.sas files somewhere whilst developing Anyone with other macro best-practice suggestions? ... View more

Just to clarify, do you wish to: Replace the stored value of missing with a stored value of zero? Keep the stored value as missing, but display it with a "0" rather than a "."? A stored value of 0 (displayed as a 0) and a missing value displayed as a 0 are not the same thing. To display missing stored values as 0, see post #7. To replace text of ( . ) with text of ( 0 ), see post #6.      data new(drop=i);           set old;           /* The following replaces stored numeric missings with stored values of 0 */           array  MyNums    _NUMERIC_ ;           do i = 1 to dim(MyNums);                MyNums=coalesce( MyNums , 0 );           end;           /* The following replaces all occurrences of one character string with another */          array  MyChars    _CHARACTER_ ;           do i = 1 to dim(MyChars);                MyChars=tranwrd( MyChars,  "( . )"   /* Old text */  ,"( 0 )"  /* New Text */  );           end;      run; BTW, I cannot tell if your displays of missing contain one or two blanks either side of the period. It appears to be 2 on the left hand side, 1 on the right hand side? ... View more

Scott - do they need to be pre-assigned? ... View more

Hi Scott - try this. It should show all currently assigned Oracle libraries: proc sql;   select * from dictionary.libnames where find(engine,"ORACLE","I")   union   select * from dictionary.libnames where find(sysdesc,"ORACLE","I"); quit; ... View more

An alternative to the "pipe" statement (which spawns an O/S shell, something not always allowed) is to use the SAS file & directory functions - dopen(), fopen(), etc - to get the list of Excel spreadsheets in a directory, each of which can then be used in a subsequent import/libname statement.  (This also removes the O/S dependency.) ... View more