I have a dataset that is read in from a control file that lists visit IDs, names, and windows as well as anchor datasets if applicable.  I then merge this with eCRF data and external lab data to compare visit dates to make sure they are within windows.  My concern is that there may be more than one anchor dataset, and I want to be able to get a distinct list of anchor datasets where applicable, but I'm not sure how to go about this.  Below is my current thought process, but this is a specific example.  I need to generalize my code inside a macro to fit any number of anchor datasets.  I think my biggest concern is how to get the maximum number of "words" in the ANCHOR variable so that I can loop through it.  There may even be a more efficient way to do it than what I have below.   In short, below is the code to get a sample dataset.  I want to take the ANCHOR variable and get a list of distinct values, which in this case would be RD_DOSEL, RD_DOSEL2, RD_DOSEL3, RD_DOSEL4, RD_DOSEL5.  I can then use this to loop inside a macro to read each dataset.     data viswin; input VISITID 1-5 VISITMNEMONIC $ 9-24 MIN_WINDOW 25-27 MAX_WINDOW 29 ANCHOR $ 33-63; cards; 30101 Day 1 (Part 1) 0 0 n/a 30115 Day 15 (Pat 1) -2 2 n/a 30199 SFUP (Part 1) -7 7 RD_DOSEL, RD_DOSEL2, RD_DOSEL3 30201 Day 1 (Part 2) 0 0 n/a 30215 Day 15 (Part 2) -2 2 n/a 30299 SFUP (Part 2) -7 7 RD_DOSEL, RD_DOSEL4 30301 Day 1 (Part 3) 0 0 n/a 30315 Day 15 (Part 3) -3 3 n/a 30399 SFUP (Part 3) -7 7 RD_DOSEL, RD_DOSEL5 ; run; data anchor; set viswin (where=(anchor^='n/a')); ndset=countw(anchor); dset1=scan(anchor,1); dset2=scan(anchor,2); dset3=scan(anchor,3); run; data anchor2; set anchor (keep=dset1 rename=(dset1=dset)) anchor (keep=dset2 rename=(dset2=dset)) anchor (keep=dset3 rename=(dset3=dset)); if dset='' then delete; run; proc sort data=anchor2 nodupkey; by dset; run;   ... View more

I have a list of adverse events (AEs), and I am supposed to find if a subject has any overlapping terms and output them into a report.  I have been able to do so, but on rare occasions, I have a term that is not overlapping in some places but is in others and is outputting all records.  I would appreciate some help finding a way to eliminate the records that are not truly overlapping.   For example, Subject 101-002 is showing that "Urinary tract infection" is overlapping, which is true for AESPID 2 & 4; however, AESPID=1 is not overlapping.  I want to get rid of this.   SUBJECTNUMBERSTR AESPID LLT_NAME AESTDAT AEENDAT 101-002 1 Urinary tract infection 03AUG2017 15AUG2017 101-002 2 Urinary tract infection 07SEP2017   101-002 4 Urinary tract infection 03OCT2017      My method for getting to this point is: Get a list of all Subject/LLT records where there is a repeat of LLT_NAME Compare all AESTDAT & AEENDAT values to see if there is an overlap and flag Keep distinct Subject/LLT records with a flag Merge the overlapping records with the original data and only keep overlapping terms  Here is the code I have.  Please feel free to modify to help me remove AESPID=1   data AEALL; input @1 SUBJECTNUMBERSTR $7. @9 AESPID @13 LLT_NAME $30. @45 AESTDAT date9. @57 AEENDAT date9.; format AESTDAT AEENDAT date9.; cards; 101-001 1 Acute respiratory failure 05AUG2017 . 101-001 2 Pleural effusion 05AUG2017 . 101-001 3 Pneumonia 07AUG2017 14AUG2017 101-001 4 Septic shock 16AUG2017 22AUG2017 101-001 5 Pneumothorax traumatic 19AUG2017 06SEP2017 101-001 6 Anemia 20AUG2017 . 101-001 7 Pneumothorax traumatic 28AUG2017 09SEP2017 101-001 8 Pneumonia 29AUG2017 08SEP2017 101-001 9 Hypotension 04SEP2017 07SEP2017 101-001 10 Hypernatremia 12AUG2017 23AUG2017 101-001 11 Thrombocytosis 15AUG2017 20AUG2017 101-001 12 Paroxysmal atrial fibrillation 13SEP2017 . 101-001 13 Respiratory infection 15SEP2017 . 101-002 1 Urinary tract infection 03AUG2017 15AUG2017 101-002 2 Urinary tract infection 07SEP2017 . 101-002 3 Pressure ulcer 11SEP2017 . 101-002 4 Urinary tract infection 03OCT2017 . ; run; proc sort data=aeall; by subjectnumberstr llt_name aestdat aeendat; where llt_name^=''; run; data duplicate; set aeall; by subjectnumberstr llt_name; if first.llt_name & last.llt_name then delete; run; proc sort data=duplicate; by subjectnumberstr llt_name aespid; run; data overlap; set duplicate; by subjectnumberstr llt_name aespid; format prev_stdat prev_endat date9.; prev_stdat=lag(aestdat); prev_endat=lag(aeendat); if first.llt_name then do; prev_stdat=.; prev_endat=.; end; else do; if (prev_endat^=. & aestdat^=. & prev_stdat<=aestdat & aestdat<=prev_endat) or (prev_endat=. & aestdat^=. & prev_stdat<=aestdat) or (prev_endat^=. & aestdat^=. & aestdat=prev_endat+1) or (aestdat=.) then flag=1; end; run; proc sort data=overlap out=overlap_distinct (keep=subjectnumberstr llt_name) nodupkey; by subjectnumberstr llt_name; where flag=1; run; data final; merge aeall overlap_distinct (in=a); by subjectnumberstr llt_name; if a; run; proc sort data=final nodupkey; by subjectnumberstr LLT_NAME AESPID AESTDAT AEENDAT; run;   ... View more

Okay.  This is getting on my nerves.  Just when I think I know how to access macro variables with multiple ampersands, something goes wrong.  Can someone correct my code so I can display the values 7 and 624?  I have tried all kinds of combinations of ampersands and periods, but the closest I can get it to display "BRON_BRON_R1_TERM_CNT" and "CM_TERM_CNT" instead of "7" and "624".   %let dsname1=BRON_BRON_R1; %let dsname2=CM; %let BRON_BRON_R1_TERM_CNT=7; %let CM_TERM_CTN=624; %macro loop; %do i=1 %to 2; %put &&&&dsname&i.._TERM_CNT; %end; %mend loop; %loop; ... View more