I also find one paper (link) which explained usage of binary variable which quote "In addition, any numeric variable without a usage of ORDER, GROUP or DISPLAY has a default usage of SUM...and that means that the correct reference for VAR4 (and for your NOPRINT items) is variable_name.SUM in the COMPUTE block." https://communities.sas.com/t5/SAS-Programming/PROC-REPORT-SAS-log-reporting-dataset-variables-are/td-p/258959 ... View more

Hi Reeza,   As you have said, I've created ab binary variable ind = ( pct_chg_lst < &pct_threshold.); where &pct_threshold =5 ( say 5% is my threshold. Now, in my proc report when I'm trying to use this binary variable , I am getting variable uninitialized message. define BASE_C_LAST / " % Chg #" style={cellwidth= 0.60 in}; COMPUTE BASE_C_LAST; if IND= 0 THEN CALL DEFINE("BASE_C_LAST", "style", "STYLE=[BACKGROUND=LAVENDERBLUSH]"); ENDCOMP; I am trying to change color of variable BASE_C_LAST based on the value of IND.    ... View more

@BrunoMueller Hi Bruno, Thanks for the solution, but what if i need like this   data have;     infile cards missover;     input id flag switch $8.; cards; 001 0 001 0 001 0 001 1 first 001 0 001 0 001 1 second 001 0 001 0 002 0 002 1 first 002 0 002 0 002 1 second 002 1 third ; run;   /*I need three outputs based on switches i,e.*/ /**/ /*data first will looks like:*/ data first;     input id flag switch $8.; cards; 001 0 001 0 001 0 001 1 first 002 0 002 1 first ; run; /*this is selecting observation from first to second*/ data second;     input id flag switch $8.; cards; 001 1 first 001 0 001 0 001 1 second 002 1 first 002 0 002 0 002 1 second ; run; data third;     input id flag switch $8.; cards; 002 1 second 002 1 third ; run;     please help! ... View more

@ramchinna24 Many Thanks for the help and time! . It worked  ... View more

@ramchinna24  Hi I have tried this:   data test;     set new_have;/*sorted on pnr date*/     by pnr;        if lag(type)='Psoriasis' and type in ('Nonbiologic_systemic' 'Biologics')   then f=1;        if lag(type)='Psoriasis' and type='Psoriasis' and lag(atc_code) ne atc_code then f=1;                    if lag(type)='Nonbiologic_systemic' and type in ('Biologics') then f=1;        if lag(type)='Nonbiologic_systemic' and type='Nonbiologic_systemic' and lag(atc_code) ne atc_code then f=1;     if lag(type) in ('Nonbiologic_systemic' 'Psoriasis') and type='Biologics' then f=1;     if lag(type)='Biologics' and type='Biologics' and lag(atc_code) ne atc_code then f=1;      /*     if lag(pnr) =pnr and lag(atc_code) ne atc_code then f=1; */ run;     But it's ignoring the by group and adding flagging  for situations when patient by group is changes. ... View more

@ramchinna24 I think, sorting order is conflicting with data, for example on: PNR 478, Switch was recorded at 5th March 2010 , Nonbiologic_systemic 2/10/2010 Psoriasis 2/23/2010 Nonbiologic_systemic 3/5/2010   but when we are sorting it on atc_code and then type and followed by edatum , switch is recording at 2/10/2010.  ... View more

@ramchinna24 Hi , code is working on treatment Type level, however whenever atc_codes are changing withing the treatment type, it's not working. I have added the desired output in previous reply, please kindly check. Many thanks for your's help!! ... View more

@ramchinna24 Hi, Thanks for the logic, but what if ATC code withing the psorasis or within other treatment group changes, we also need to identify those changes as switches.   i want this : lpnr ATC_CODE TYPE EDATUM Flag_switch Comment 478 L04AX03 Nonbiologic_systemic 2/10/2010     478 D05AX52 Psoriasis 2/23/2010     478 L04AX03 Nonbiologic_systemic 3/5/2010 Y This is switch because subject is changing his treatment from lower(psorais) to higher treatment(non_bio_sys/bio) 478 L04AX03 Nonbiologic_systemic 5/18/2010     478 L04AX03 Nonbiologic_systemic 8/2/2010     478 L04AX03 Nonbiologic_systemic 11/3/2010     478 L04AX03 Nonbiologic_systemic 2/7/2011     478 L04AX03 Nonbiologic_systemic 8/16/2011     478 L04AX03 Nonbiologic_systemic 11/22/2011     603 D05AX02 Psoriasis 8/24/2005     603 D05AX02 Psoriasis 10/13/2005     603 D05AX02 Psoriasis 6/2/2006     603 D05AX02 Psoriasis 7/19/2006     603 D05AX02 Psoriasis 4/3/2007     603 D05AX02 Psoriasis 12/7/2007     603 D05AX02 Psoriasis 1/4/2008     603 D05AX02 Psoriasis 10/2/2008     603 D05AX52 Psoriasis 3/16/2009 Y This is switch because subject is changing his ATC code within Psorasis(or any other treatment group) 603 D05AX52 Psoriasis 5/7/2009     603 D05AX52 Psoriasis 8/21/2009     603 D05AX52 Psoriasis 11/9/2009     603 D05AX52 Psoriasis 3/19/2010     603 D05AX52 Psoriasis 7/30/2010     603 D05AX52 Psoriasis 1/13/2011     603 D05AX52 Psoriasis 5/3/2011     603 D05AX52 Psoriasis 12/12/2011     1103 L04AX03 Nonbiologic_systemic 1/17/2006     1103 L04AX03 Nonbiologic_systemic 8/23/2006     1103 L04AX03 Nonbiologic_systemic 11/9/2006     1103 L04AX03 Nonbiologic_systemic 4/19/2007     1103 D05AX52 Psoriasis 10/26/2007     1103 D05AX52 Psoriasis 4/1/2008     1103 L04AX03 Nonbiologic_systemic 10/20/2008 Y This is switch because subject is changing his treatment from lower(psorais) to higher treatment(non_bio_sys/bio) 1103 D05AX52 Psoriasis 10/20/2008     1103 L04AX03 Nonbiologic_systemic 4/7/2009 Y This is switch because subject is changing his treatment from lower(psorais) to higher treatment(non_bio_sys/bio) 1103 L04AX03 Nonbiologic_systemic 11/20/2009     1103 L04AX03 Nonbiologic_systemic 11/12/2010     1103 L04AX03 Nonbiologic_systemic 8/12/2011     1103 D05AX52 Psoriasis 10/22/2011     1103 D05AX52 Psoriasis 12/29/2011     ... View more

@DartRodrigo Hi , Thanks for the quick reply. I am sorry for the unclear explaination, but what i exactly means is: as you see the data, we have three treatment type in TYPE variable. Now for each Type variable we have different ATC_CODEs. I need to identify the event (SWITCH) , whenever subject changes his treatment type from Lower treatment type( i,e Topical) to Higher treatment group(i,e Non_bio_sys or Biologics) during his treatment period. Same is true for Non_sys_Bio , if Subject changes to Biologics, then it will also consider as Switch.  Event is also to be counted whenever subjects changes his ATC_CODEs within his treatment TYPE , ie in case, when Treatment TYPE is "Psorasis" and if Its' ATC_CODE changes , then we have to identify it as SWITCH.  same is true for other treatment arm.   order of treatment: BIOLOGICS> NON_BIO_SYS> TOPICAL.   Hope i am able to clear you doubt. please help. thanks ... View more