For GTL, this would be in the TICKVALUEATTRS=(family="<some font family>") option inside of the appropriate axis opts (e.g. XAXISOPTS, YAXISOPTS, etc.). For the SGPLOT and SGPANEL, this would be VALUEATTRS=(family="<some font family>") on the appropriate axis statement (e.g. XAXIS, YAXIS, etc.). ... View more

Use the RANGE option to specify the upper left corner and PROC IMPORT will import that. So to skip the first row then start in A2.  To skip 4 lines start in A5. proc import file='c:\downloads\test_range.xlsx' dbms=xlsx out=want replace; range='$A2:' ; run; If the file also includes junk below or to the right of the table then specify the full range. proc import file='c:\downloads\test_range.xlsx' dbms=xlsx out=want replace; range='$A2:E18' ; run;   ... View more

Post usable example data in a data step with datalines, as you have been shown here and as you have done yourself here. ... View more

Works with the data you posted. Let's add two variables to your date.  One is a binary flag EXPECT to indicate which observations should be "kept". The other is a record counter so we can force SQL to output the observations in the same order as they came in. data lb; input subjid testcd $3. blfl :$1. expect ; row+1; cards; 1 ALB . 1 1 ALB . 1 1 ALB . 1 1 ALB Y 1 1 ALB Y 1 1 ALP . 0 1 ALP . 0 1 ALP Y 0 2 ALB . 0 2 ALB . 0 2 ALB . 0 2 ALB Y 0 2 ALP Y 1 2 ALP Y 1 3 ALB Y 0 3 ALP Y 1 3 ALP Y 1 3 ALP Y 1 4 ALB . 0 4 ALB . 0 4 ALB . 0 4 ALB Y 0 4 ALP . 1 4 ALP . 1 4 ALP . 1 4 ALP Y 1 4 ALP Y 1 4 ALP Y 1 4 ALT Y 1 4 ALT Y 1 ; proc sql; create table test1 as select * from lb group by 1,2 having count(blfl) > 1 order by row; create table test2 as select * from lb group by 1,2 having sum('Y'=blfl) > 1 order by row; quit; proc compare data=test1 compare=test2; id row; run; proc compare data=test1 compare=lb(where=(expect)); id row ; run;   ... View more

Your SQL code does not work with your sample data.  If you could supply SQL code that works properly with the data (and data which has numeric variables where the SQL expects them), then it will be much easier to determine the equivalent DATA step code.   The problem you have identified, doing a many-to-many in a DATA step takes a bit of coding.  But use of a hash object can make it reliable and useful. ... View more

Can you post the output you want to see ? ... View more

Why does your sample expected output include a subject=5 and aedecod="Blood creatinene increased"?   That output record is the only input record for this subject/aedecod combination.  Yet your stated criterion includes "subjects that have same AEDECOD".  I take this to mean records in which a given subject/aedecod combination appears more than once (and then keep those records which also have a non-missing date and AEOUT="Not-recovered/Not-resolved").  If your sample output is not a mistake, could you make your objective more precise?  I'm missing something.   And I have the same question about why your output has a subject=5 and aedecod="Chloroma".  Again, "Chloroma" appears only once for this subject.   Now if those records should not be in the output, then this program does what I understand you want.   data have; x=_n_; input subject @5 aedecod $ 5-42 @43 aestdtn yymmdd10. @54 aeout $ 54-80 @81 aeendtn yymmdd10.; *format aestdtc aeendtc yymmdd10.; format aestdtn aeendtn yymmdd10.; datalines; 001 Fatigue 2019-06-03 Not-recovered/Not-resolved . 001 Nausea 2019-06-03 Not-recovered/Not-resolved . 001 Vomiting 2019-06-03 Not-recovered/Not-resolved . 003 Asthenia 2019-08-28 Recovered/Resolved 2019-08-28 003 Hypothension 2019-06-25 Recovered/Resolved 2019-08-08 003 Thrombocytopenia 2019-07-11 Recovered/Resolved 2019-07-23 003 Thrombocytopenia 2019-07-18 Recovered/Resolved 2019-08-04 003 Thrombocytopenia 2019-08-12 Not-recovered/Not-resolved 2019-08-12 004 Anaemia 2019-07-10 Not-recovered/Not-resolved . 004 Blood creatinine increased 2019-08-19 Recovered/Resolved 2019-09-04 004 Blood creatinine increased 2019-09-02 Recovered/Resolved 2019-11-13 004 Blood creatinine increased 2019-11-11 Recovered/Resolved 2019-12-08 004 Epistaxis 2019-01-21 Not-recovered/Not-resolved . 004 Epistaxis 2019-11-25 Recovered/Resolved 2019-11-28 004 Thrombocytopenia 2019-12-02 Not-recovered/Not-resolved . 004 Upper respiratory tract infection 2019-10-10 Recovered/Resolved 2019-10-13 005 Alanine aminotransfrerase increased 2019-08-14 Not-recovered/Not-resolved 2020-08-20 005 Alanine aminotransfrerase increased 2019-08-21 Recovered/Resolved 2020-09-06 005 Alanine aminotransfrerase increased 2019-09-07 Not-recovered/Not-resolved . 005 Aspartate aminotransfrerase increased 2019-08-14 Not-recovered/Not-resolved 2020-08-20 005 Aspartate aminotransfrerase increased 2019-08-21 Recovered/Resolved 2020-09-06 005 Aspartate aminotransfrerase increased 2019-09-07 Not-recovered/Not-resolved . 005 Blood creatinine increased 2019-05-29 Not-recovered/Not-resolved 2020-07-20 005 Cellulitis 2020-07-01 Not-recovered/Not-resolved 2020-08-19 005 Cellulitis 2020-08-20 Recovering/Resolving 2020-09-25 005 Cellulitis 2020-08-26 Not-recovered/Not-resolved 2020-10-01 005 Chills 2019-12-26 Recovered/Resolved 2019-12-27 005 Chloroma 2020-04-30 Not-recovered/Not-resolved 2020-12-30 005 Cholelithiasis 2020-08-18 Recovered/Resolved 2020-08-20 005 Epistaxis 2019-12-17 Recovering/Resolving 2019-12-26 005 Pyrexia 2020-08-16 Recovered/Resolved 2020-08-20 005 Pyrexia 2020-08-20 Not-recovered/Not-resolved . 005 Subcutaneous haematoma 2020-01-24 Not-recovered/Not-resolved . 005 Urinary Tract infection 2020-08-20 Recovering/Resolving 2020-08-25 005 Urinary Tract infection 2020-08-26 Not-recovered/Not-resolved . run; data want; set have (where=(aeout^='Not-recovered/Not-resolved')) have (where=(aeout='Not-recovered/Not-resolved' and aeendtn=.)) have (where=(aeout='Not-recovered/Not-resolved' and aeendtn^=.)); by subject aedecod; if aeout='Not-recovered/Not-resolved' and aeendtn^=.; /*all qualifying records ...*/ if not (first.aedecod=1 and last.aedecod=1) ; /* ... that are not singletons */ run;   Fortunately, your data are already sorted by subject/aedecod.  So for each subject/aedecod the program above first reads all instances that do not have aedecod="Not-recovered/Not-resolved", followed by any dateless instances that do have aedecod="Not-recovered/Not-resolved, finally followed by qualifying records (non-missing date and aedecod="Not-recovered/Not-resolved").  So as long as those qualifying records are not the only ones for that subject/aedecod, the record is output. ... View more

@RAVI2000 wrote: proc transpose data = have out=want prefix=_; by trt subject test descending grade; var grade; id test; run; That's not the code I provided. In fact the code I provided seems to work perfectly on your data. ... View more

@RAVI2000 wrote: I would like to see how race_ethnicity is effecting the RPL_Themes If you stop here, then GLM is correct. But then you add   ... and how age is effecting the RPL_Themes.   and so now maybe there's some confusion over words. Correlation does not imply causation, and so you could compute some sort of correlation between these two variables except its not clear again how you are meaning the word "effecting" which maybe should say "affecting" but even there, its not clear what you mean.   If you want to quantify relationships between the two variables, that's possible, but you haven't said that's what you want. If you want to find predictive relationships, then we're back to GLM and that is really only useful in the case where one variable predicts the other; and not in the case where you seem to be asking for where variable 1 predicts variable 2 and also at the same time variable 2 predicts variable 1. ... View more

Looks like your multiple cars are nicely separated by commas. So this may work: data want; set have; do _N_=1 by 1; car=scan(cars,_N_,','); if car=' ' then leave; output; end; drop cars; run; ... View more