If I'm having someone else review the log, it's easier to make comments, etc in Word. ... View more

Ok, let's try it this way.   This code gives me the desired result but I have to use a data step and then sql.   I think I should be able to do it all in sql.  However, if this method is the only way to do it, than I can stick with it.   /*Create dataset*/ data a; input x 1. y mmddyy10.; format y mmddyy10.; list; datalines; 111-02-23 111-05-23 112-01-23 112-25-23 211-01-23 211-09-23 311-05-23 311-07-23 311-11-23 ; run;   /*group observations with dates between 0 and 5 days apart*/ data b(drop=y_lag); set a; by x y; y_lag=lag(y); if first.x or y gt (y_lag + 5) then cnt+1; run;   /*This gives me the desired result*/ proc sql; create table c as select x, y from b group by x, cnt having count(*) gt 1; quit; ... View more

Only comparing as part of the same group so only the 2nd and 3rd rows in your example would qualify. ... View more

No, not just looking at pairs, I added a few more rows to illustrate.  In this dataset, the 1st and 2nd obs would qualify, so would the 7th, 8th, and 9th obs   27 /*Create dataset*/ 28 data a; 29 input x 1. y mmddyy10.; 30 format y mmddyy10.; 31 list; 32 datalines; RULE: ----+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8----+----9----+----0 33 111-02-23 34 111-05-23 35 112-01-23 36 112-25-23 37 211-01-23 38 211-09-23 39 311-05-23 40 311-07-23 41 311-11-23 NOTE: The data set WORK.A has 9 observations and 2 variables. NOTE: DATA statement used (Total process time): real time 0.01 seconds cpu time 0.01 seconds 42 ; 43 run; 44 45 /*print contents in log*/ 46 data null; 47 set a; 48 put x= y=; 49 run; x=1 y=11/02/2023 x=1 y=11/05/2023 x=1 y=12/01/2023 x=1 y=12/25/2023 x=2 y=11/01/2023 x=2 y=11/09/2023 x=3 y=11/05/2023 x=3 y=11/07/2023 x=3 y=11/11/2023 NOTE: There were 9 observations read from the data set WORK.A. NOTE: The data set WORK.NULL has 9 observations and 2 variables. NOTE: DATA statement used (Total process time): real time 0.00 seconds cpu time 0.01 seconds ... View more

Ok, see if this is clearer, I put date values in y.   I'd like to select observations that have the same value of X and their date values (Y) are within 5 days of each other.   The first and second observations would be selected because have the same value of X and are within 3 days of each other.   The third and fourth observations have the same value of X but would not be selected because their date values after 8 days apart.   28 /*Create dataset*/ 29 data a; 30 input x 1. y mmddyy10.; 31 format y mmddyy10.; 32 list; 33 datalines; RULE: ----+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8----+----9----+----0 34 111-02-23 35 111-05-23 36 211-01-23 37 211-09-23 38 311-05-23 NOTE: The data set WORK.A has 5 observations and 2 variables. NOTE: DATA statement used (Total process time): real time 0.01 seconds cpu time 0.01 seconds 39 ; 40 run; 41 42 /*print contents in log*/ 43 data null; 44 set a; 45 put x= y=; 46 run; x=1 y=11/02/2023 x=1 y=11/05/2023 x=2 y=11/01/2023 x=2 y=11/09/2023 x=3 y=11/05/2023 NOTE: There were 5 observations read from the data set WORK.A. NOTE: The data set WORK.NULL has 5 observations and 2 variables. NOTE: DATA statement used (Total process time): real time 0.01 seconds cpu time 0.01 seconds ... View more

Actually, the fmtinfo does give the format type (see my example in the first post on this topic).    I'm just looking for a way to generate the result without having to run a data step. ... View more

Sorry, it will give the type as "num" or "char" but not "date" ... View more

Dictionary.columns will give me the format itself, i.e. yymmdd10. but it won't give me the type (num, char, date, etc). ... View more

ok, I see the problem.   It looks like I'm getting duplicate bene_sk values whenever there is a match on the first and third data sets.   I've printed the first several rows of the duplicate bene_sk cases below.  I can't print that variable in this forum, but each two rows has same value and they alternate between having a value in lomajor or upmajor.   Is there a place I can insert a statement of "on a.bene_sk=c.bene_sk"?     84 proc sql; 2 The SAS System 09:42 Friday, August 25, 2023 85 create table d as 86 select * 87 from tbl2 88 group by bene_sk 89 having count(*) gt 1; NOTE: The query requires remerging summary statistics back with the original data. NOTE: Table WORK.D created, with 58 rows and 4 columns. 90 quit; NOTE: PROCEDURE SQL used (Total process time): real time 0.03 seconds cpu time 0.01 seconds 91 92 data _null_; 93 set d; 94 put lomajor lominor upmajor; 95 run; 1 . . . . 1 . . 1 1 . . . . 1 1 . . . . 1 1 . . 1 . . . . 1 . . 1 1 . . . . 1 1 . . ... View more