It took me a while to understand what you want, but I think you want to count each time a date is more than two days from an earlier date? If this is the case, then you want to count this in "Variable"? You can achieve this as follows. data input; attrib ID length=8 Date length=8 informat=mmddyy10. format=date9. Variable length=8 ; infile datalines delimiter='09'x missover dsd; input ID Date Variable; datalines; 1 8/28/2012 0 1 8/30/2012 0 1 11/5/2012 1 1 1/2/2013 2 1 2/13/2013 3 1 2/13/2013 3 1 2/14/2013 3 1 2/15/2013 3 1 2/15/2013 3 1 2/17/2013 4 1 3/18/2013 5 1 4/15/2013 0 2 5/1/2008 0 2 5/20/2008 1 2 6/14/2008 2 ; run; proc sort data=input; by ID Date; run; data want; set input; by ID; * Retain value from one observation to the other. ; retain Variable_Calculated Counted_Date; * Reset if we start with a new ID. ; if first.ID then do; Variable_Calculated = 0; Counted_Date = Date; end; else if Date - Counted_Date > 2 then do; Variable_Calculated + 1; Counted_Date = Date; end; format Counted_Date date9.; run; proc print data=want; run; ID Date Variable Variable_Calculated Counted_Date 1 28AUG2012 0 0 28AUG2012 1 30AUG2012 0 0 28AUG2012 1 05NOV2012 1 1 05NOV2012 1 02JAN2013 2 2 02JAN2013 1 13FEB2013 3 3 13FEB2013 1 13FEB2013 3 3 13FEB2013 1 14FEB2013 3 3 13FEB2013 1 15FEB2013 3 3 13FEB2013 1 15FEB2013 3 3 13FEB2013 1 17FEB2013 4 4 17FEB2013 1 18MAR2013 5 5 18MAR2013 1 15APR2013 0 6 15APR2013 2 01MAY2008 0 0 01MAY2008 2 20MAY2008 1 1 20MAY2008 2 14JUN2008 2 2 14JUN2008 Based on this, I would say that your value of "Variable" for 15 April 2013 for ID 1 is wrong. Should that not be 6 instead of 0? 15 April is more than 2 days apart from 18 March.   By the way, I am relying on the internal representation of dates in SAS, which is the number of days since 1 January 1960. This means I can just subtract date variables if I am interested in the number of days. ... View more

Try this   data have; input ID $ Number; datalines; A 1 A 2 A 5 A 19 A 19 B 1 B 3 ; data want; set have; by ID; if dif(Number) not in (0, 1) then numberoftimes + 1; if first.ID then numberoftimes = 0; run;     Result:   ID Number numberoftimes A 1 0 A 2 0 A 5 1 A 19 2 A 19 2 B 1 0 B 3 1 ... View more

Hi, @Emma8,   Are you asking for help with SAS code or with Python code?  I think you're asking re Python, yes?  It may be a bit more difficult to get Python code in a SAS coding forum.  Just a thought.   Jim ... View more

data have; infile cards dsd; input ID $ Datetime :datetime20.;* NEW_ID; format datetime datetime20.; cards; AAC56,25AUG20:02:00:00,1 AAC56,27AUG20:21:00:00,2 AAC56,31AUG20:10:30:00,3 AAC56,02SEP20:20:00:00,4 AAC56,29SEP20:21:43:00,5 AAC56,29SEP20:21:43:00,5 AAC56,29SEP20:22:03:00,5 AAC56,29SEP20:23:03:00,5 AAC56,30SEP20:00:03:00,5 AAC56,30SEP20:01:03:00,5 AAC58,28SEP20:13:45:00,1 AAC58,29SEP20:15:00:00,2 AAC58,29SEP20:16:00:00,2 AAC58,29SEP20:16:00:00,2 AAC58,29SEP20:17:10:00,2 ; %let hour=1;/*Assign your interval here i.e. 1 or 2 etc*/ data want; do until(last.id); set have; by id; if first.id=1 or intck('hour',n,datetime)>&hour then new_id=sum(new_id,1); n=datetime; output; end; drop n; run; ... View more

Hi @Emma8  Thank you. Please try- data have; infile cards dsd; input ID $ Datetime :datetime20.;* NEW_ID; format datetime datetime20.; cards; AAC56,25AUG20:02:00:00,1 AAC56,27AUG20:21:00:00,2 AAC56,31AUG20:10:30:00,3 AAC56,02SEP20:20:00:00,4 AAC56,29SEP20:21:43:00,5 AAC56,29SEP20:21:43:00,5 AAC56,29SEP20:22:03:00,5 AAC56,29SEP20:23:03:00,5 AAC56,30SEP20:00:03:00,5 AAC56,30SEP20:01:03:00,5 AAC58,28SEP20:13:45:00,1 AAC58,29SEP20:15:00:00,2 AAC58,29SEP20:16:00:00,2 AAC58,29SEP20:16:00:00,2 AAC58,29SEP20:17:10:00,2 ; data want; do until(last.id); set have; by id; if first.id or intck('hour',_n_,datetime)>1 then new_id=sum(new_id,1); _n_=datetime; output; end; run; Fyi- The correction is- if first.id or intck('hour',_n_,datetime)>1 then new_id=sum(new_id,1); _n_=datetime; ... View more

 My typo - code for &ID with more than one varaible should be:       if varname not in ("&IDS") then output; As you defined only one variable variable: subid the code need be  changed into:      if varname not = "&IDS"  then output;   ... View more

data test; set sashelp.class; if age in (12,13) then gr=1; else if age =14 then gr=2; else gr=3; run; proc sort data=test; by gr; run; proc format; value gr 1="kids Reference" 2="teenagers 9 (4-24)" 3="old 78 (32-189)" ; proc template; define style mystyle; parent=styles.sasweb; class graphwalls / frameborder=on; class graphbackground / color=white; end; run; ods graphics on / reset=all; ods html style=mystyle path='.' file='newgraph.html'; ods graphics / reset=all width=6.5in height=4.5in border=off ; legend1 label=none; proc sgplot data=test DATTRMAP=attrmap1; scatter x=weight y=Height /group=gr ; keylegend / location=outside position=bottom NOBORDER down=4; label gr='Age, years OR (95% CI)'; format gr gr.; run; The least you can do is: 1. Please format the code you post 2. Please run the code you post; This code results in an error.   Also: 3.The only way to align text in the manner you want is to use fixed-pitch font, as you have been told. Do you understand this? 4. If you do, you can see why this aligns the legend as you request: proc format; value gr 1="kids Reference" 2="teenagers 9 (4-24)" 3="old 78 (32-189)" ; ... proc sgplot data=TEST; ... keylegend / ... valueattrs=(family='Courier New') ; ... run;   Here is the image (broken on some browsers it seems):               ... View more