Assuming that you mean all the program code you want is saved in separate program code files (with .sas suffix), the easiest way would be to open a new project and import all the program code files you want. Richard ... View more

Transposing the data allows all the incidents per ID to be in the same row, so they are all available in memory.  The un-transposed data as shown could be analysed with multiple lag functions but you would need to know the maximum number of events per ID for the period to set up the required number of lags.  You say you have a solution for the transposed data so I would stick with that (kinda curious to see how you proceeded, though). Richard ... View more

Should mention that the single quote does not work for Apple OSX version of MS Excel 2008. Richard ... View more

A variant on RW9's second alternative is to use a prefixed single quote, which at least up to Excel 2007 is recognised as a character flag (for compatibility with good old Lotus123).  The quote may appear initially when opened in Excel but it should subsequently be masked. data have;      Length ID $<one more than in original data> ;      set have;      ID="'"||strip(id);      ... run; then export that. Richard ... View more

Maisha The method I outlined will work if those 'other' variables are null except where the relevant in_sourcefilen = 1. However, there is probably a better way to merge the data in the first place, if you still have access to the original source files.  Assuming they are all sorted by ID: Data want ;      Length    ID  in_sourcefile1 - in_sourcefileN      8 ;      Merge     file1 (in = in_file1 keep = ID <list of variables to keep> )                     file2 (in = In_file2 keep = ID <list of variables to keep> )                     ...                     ;      By     ID ;      Array in_source {*} in_sourcefile: ;     /* note semicolon modifier */      Array in_file {*} in_file: ;                     /* note semicolon modifier */      Do k = 1 to DIM (in_source) ;           If in_file {k} then in_source {k} = 1 ;      End ; Run ; If there is no overlap between the variables in each of the source files, other than ID, you can omit the keep = options. Richard ... View more

If the following conditions are met in your data - values of in_sourcefilen are either 1 or null (not zero) - the most up to date values of the 'other' variables is in the last (or in all) record for each ID then the following approach might work: Proc sort equals /* preserves relative order within id */           data = have           out = step1           ;      By     ID ; Run ; /* create master dataset with unique ids */ Data want ;      Set step1 (Keep = ID) ;      By ID ;      If First.ID ; Run ; /* use update to add the remaining data */ data want ;      update want                step1                ;      By ID ; Run ; Warning - untested code Even if it works, examine closely to ensure this is the result you want. Richard ... View more

And here is a way to change your code to date literals.  Simply paste all your if statements between Datalines4 and ;;;; and run - then copy the new if condition from the log and paste them into your original code.  This only has to be done once. Warning - untested code!  So make sure you have a backup copy of your original code.  Data _NULL_ ;   Infile Datalines ;   Input ;   Length Date_Lit $ 12 ;   p1 = find("=", _Infile_, 20) ; /* Second = */   p2 = find(";", _Infile_) ; /* First ; */   Part_A = Substr(_Infile_, 1, p1) ;   Part_C = Substr(_Infile_, p2, 20) ;   Date_MMDDYY = Substr(_Infile_, p1+1, p2-p1-1) ;   Date_Date9 = Put(Input(Date_MMDDYY, MMDDYY10.), Date9.) ;   Part_B = catt("'", Date_Date9, "'D") ;   Result = catt(Part_A, Part_B, Part_C) ;   Put Result ; Datalines4 ; if var_a=1 then date01=01/1/2012; else date01= .; if var_b=1 then date02=03/15/2013; else date02=.; ... ;;;; Hope this works Richard ... View more

You are correct that the expression '01/01/2012'D will not be interpreted as a SAS date.  You could adopt pradeepalankar's solution by copying and pasting the input("  and ",mmddyy10.) bits before and after all your dates.  Be aware that this is a less efficient solution for a large dataset because the input function needs to be evaluated for each condition each time, whereas the date literals are evaluated only once.  If you have more than a million rows you will notice the difference. Richard ... View more

Maisha If you have correctly copied your code the problem lies with the way you have expressed the dates. Sas interprets date01=01/1/2012 to be 1 divided by 1 divided by 2012, a number close to zero.  Hence when formatted as a date it is shown as zero days from 1 Jan 1960, the 'zero' point for SAS dates. You need to use a SAS date literal, which is a DATE7., DATE9. or DATE11. expression enclosed in quotes and followed by a D. Try if var_a=1 then date01='01JAN2012'D ; else date01= .; Richard ... View more

It sounds like you have individual triggers on each of the 16 input file (if not, I think you should  have) and an AND condition in LSF which requires all conditions to be met.  If you could insert a step after the AND condition is met that creates a trigger for the rest of the process if and only if the the date is the same as the initial date of the process (which could be a macro variable set by a step triggered by the 6am start) then this would ensure that the process would halt if all the trigger conditions were not met by midnight, and depending on your setup, move the input files out of the way.  The next day a fresh process would be initiated un-compromised by files left over from the previous day.  The halted processes would need to be killed but they would provide diagnostic information. Richard ... View more

Proc Freq does not appear to have the flexibility you require to preserve a particular order in summary reporting.  Tabulate, Means and Report all allow ordering by a prearranged format.  To implement tis you need to create the format using the NOTSORTED option , and then report specifying both ORDER=DATA and PRELOADFMT options. See  Base SAS(R) 9.2 Procedures Guide I would suggest creating the healthhad format as Proc Format ;      Value $healthhad NOTSORTED           'q101r1' = 'Private insurance that is provided by your current or former employer or union'           'q101r2' = ' Insurance that is provided through a family members employer or union'           [etc] Then code your data using coded values for the column helathhad Data want ;      Set have ;      format healthhad $healthhad. ;      Select ;           When q101r1 then healthhad = 'q101r1' ;           When q101r2 then healthhad =  'q101r2' ;      [etc]           Otherwise ;      end ; ... Run; Then use Proc Tabulate or Report to count and report the values. Richard ... View more

This is a common statement in many procedures , its use is to specify the name of the output dataset and any options to use during output.  See SAS/STAT(R) 9.22 User's Guide for more detail, in particular: DAPPROXIMATIONS outputs approximations to transformed dependents I would suggest you read the documentation (freely available) before using any statistical procedure. Richard ... View more

When you do have matching data you should consider rephrasing your if condition:      if x and y; This is equivalent to      if x = 1 and y = 1 ;  /* ie both true */ The condition x=y is also met when both are false, an unwanted case.  Sure, the merge process takes care of that but it is a sloppy statement in this context. Richard ... View more